Great Lakes Research Review
Volume 3, Number 1, April 1997
Great Lakes Exotic Species
What Other Ecosystem Changes Have Zebra Mussels Caused in Lake Erie: Potential Bioavailability of PCBs
Joseph V. DePinto
Rajagopal Narayanan
Great Lakes Program
University at Buffalo
Buffalo, NY 14260-4400
(716) 645-2088 fax: (716) 645-3667
Email: depinto@eng.buffalo.edu

Introduction
Lake Erie has undergone tremendous changes over the past 15-20 years. Most of those changes can be attributed to phosphorus loading control measures implemented in the basin. However, some of the recent changes may be the result of the significant invasion of zebra mussels (and more recently quagga mussels) to the lake. In any event there is considerable interest in developing an understanding of these ecosystem changes and how they are related to management actions on Lake Erie (e.g., nutrient control, toxic chemical load reduction, fish consumption advisories, fish management programs). This interest is evidenced by publications such as the Journal of Great Lakes Research special issue on "Evidence for the Restoration of Lake Erie," 19(2), 1993 and by workshops such as "The Changing Face of the Lower Great Lakes Ecosystems," co-hosted by the New York Sea Grant Institute and the Great Lakes Program at the University at Buffalo (February 5, 1994). Also, the stakeholders within the Lake Erie basin are currently in the process of developing and implementing a Lake Erie Lakewide Management Plan (LaMP), which has the task of identifying beneficial use impairments in Lake Erie as a whole and developing and implementing a management plan for eliminating those impairments. This LaMP process requires the use of an Ecosystem Approach for managing the lake and, therefore, requires a quantitative understanding of the Lake Erie ecosystem structure, function, and response to multiple stressors acting in concert.

Effective management of the Lake Erie ecosystem requires an ecosystem level understanding of certain stress-response observations that have been identified as key to the new Lake Erie ecosystem dynamics. Among the key observations are the following:

  • Over the past 20 years there has been a large reduction in phosphorus loadings to the lake. By the mid- 1980s the loading was reduced by about 15,000 metric tons from about 25,000 metric tons per year in the early '70s to about 10,000 metric tons per year (q2000 mta depending on annual runoff) since 1985. This led to a decrease in phytoplankton biomass (the base of the food chain) by approximately a factor of two between 1970 and 1987 (Makarawicz, 1993). Nicholas and Hopkins (1993) have synthesized the historical trend of annual TP loading and phytoplankton biomass in the western basin of Lake Erie from 1974-1990 (Figure 1). It demonstrates the relationship between phosphorus load reductions and phytoplankton response through the mid-80s as predicted by eutrophication models (DiToro, et al.1987). But with the zebra mussel invasion in the western basin taking hold in 1988, we can see an additional decrease in phytoplankton biomass in 1989-90, with no decrease in phosphorus load. The only logical hypothesis is that it is caused by mussels feeding on algae, which imposes an additional algal loss mechanism that must be considered in current nutrient-phytoplankton dynamics models.

    Figure 1: HIstorical trends of total phosphorus loading and phytoplankton densities in western Lake Erie (from Nicholls and Hopkins 1993).

  • Algal biomass decreases subsequent to the zebra mussel invasions in western Lake Erie and Saginaw Bay have resulted in a significantly increased water clarity. Pre- and post-zebra mussel Secchi depth measurements in Saginaw Bay and western basin of Lake Erie show obvious relationship with the zebra mussel invasion in these systems (Figure 2).

    Figure 2: Secchi Depth measurements in Saginaw Bay (from GLERL 1994) pre- and post-zebra mussel invasion in these systems

  • The decrease in particulate matter in the water column of Lake Erie, which is apparently the result of a zebra mussel- induced increased flux of particulate matter to the bottom sediments, should also increase the flux of hydrophobic organic chemicals (HOCS) such as PCBS to the sediments of the system. Measurements of PCBS in the soft tissues of zebra mussels in Lake Erie have ranged from about 0.1-0.5 g/g (wet weight) (Kreis, personal communication of results from a number of sources); therefore, there is no question that zebra mussels are processing and bioaccumulating PCBS. However, it is unlikely that their chemical assimilation efficiency is 100%, so not all of the mussel-induced PCB flux to the sediments is bioaccumulating in the mussels. Much of it is probably residing in the sediments as zebra mussel feces and pseudofeces that is potentially resuspendable and/or available for exposure to other benthic organisms.

  • Finally, there is the question of how the changes in phytoplankton levels and zebra mussel impacts on particulate matter cycling is affecting body burdens of bioaccumulative chemicals of concern (BCCs like PCBS and Hg) in sport fish like walleye or lake trout. Historically, PCB bioaccumulation in top predator fish has not been as much of a concern in Lake Erie as in Lake Ontario or Lake Michigan; however, the post-zebra mussel PCB levels in Lake Erie walleye (Figure 3) and lake trout are not declining and may even be increasing (Devault, et al. 1996; Whittle, Canada Department of Fisheries and Oceans unpublished data). It is possible that, due to the decrease in water column suspended solids concentrations, there has been a change in the phase distribution of the PCBS remaining in the water column toward a higher fraction of the dissolved (and therefore bioavailable) fraction. Thus, any decrease in total PCB concentration in the water column is compensated by a shift toward a higher fraction of the total being bioavailable.

    Figure 3: Total PCB concentration in Lake Erie walleye from 1977-1992 (from Devault et al. 1996)

    Obviously, a system-level model that simultaneously accounts for hydrology and hydraulic transport, nutrient loading and dynamics, phytoplankton growth and dynamics, zebra mussel growth and particulate matter processing, and PCB loading and cycling is needed to synthesize the above observations into a quantitative understanding of the mechanisms and process interactions at work in Lake Erie. In an effort to formulate and preliminarily test hypotheses explaining the above observations, we have developed and applied a screening-level solids-zebra mussel-PCB model for Lake Erie (DePinto, et al. 1996; DePinto, et al. 1997).

    Model Development
    The strategy of conducting a screening level modeling analysis for a system prior to development and application of a more complex modeling framework has proved to be quite successful in addressing both eutrophication (e.g., Chapra, 1977; Vollenweider, 1975) and toxic chemical problems (e.g., Thomann and DiToro, 1983; Rodgers, et al. 1987; Raghunathan, 1990; Endicott, et al., 1992) in the Great Lakes. Screening models can provide valuable insights about the behavior of a system without a large expenditure of funds to obtain the type of data set necessary to calibrate and confirm a complex modeling framework. Through the application of a screening model to already available data for a system, hypothesis formulation and sensitivity analysis can serve to:

  • synthesize what is known about the state of a system;

  • identify key data and process understanding gaps;

  • facilitate the framing of more complex management questions; and

  • optimize (by identifying necessary and unnecessary process terms in the model) the utility and reliability of more complex modeling frameworks used to address the more complex management questions.

    The conceptual diagrams for solids and PCBS in the Lake Erie ZEBRA MUSSELS/PCB screening model are presented in Figure 4. The solids mass balance (Figure 4a) contains two types of solids (sorbents) in the water column (biotic=algal biomass and abiotic=all other suspended particulate matter). It is important to distinguish solids type at least at this level because of the obvious differences in organic carbon content (and hence PCB partitioning), settling rates, sources, and zebra mussel food value between viable phytoplankton and other particulate matter. The model produces biotic solids (as measured by chlorophyll a) by a fairly simple phytoplankton primary production equation that depends on temperature and light but assumes a nutrient-saturated growth rate condition. There is a single first-order conversion of biotic solids to abiotic solids that lumps multiple processes such as algal death and decay and zooplankton grazing and excretion. Both biotic and abiotic solids "settle" to the sediments, albeit at different rates; however, once biotic solids enter the sediment compartment, they are immediately converted to the sediment abiotic pool. Therefore, there are three sources of abiotic solids in the water column: external loading, biotic solids conversion, and resuspension of abiotic solids from the bottom sediments. Net burial of sediments from the upper mixed layer closes the solids mass balance.

    Conceptual diagrams for Lake Erie Zebra MUSSELS/PCB screening model: a) solids mass balance; b) PCB mass balance

    Adding zebra mussels to the system does not change any of the previous solids dynamics, but adds an additional flux of solids from the water column due to the siphoning by mussels. This acts like an increased "settling" flux, but, of course that is not the mechanism. What the model calculates is the removal of particles from the water column in the vicinity of the zebra mussels at the bottom accompanied by a vertical transport of particles by settling, diffusion, and turbulent mixing that moves more particles downward from higher in the water column. Also, an additional sediment solids state variable for zebra mussel feces and pseudofeces has been included to account for rejected and excreted particulate material. Again, this separate compartment is included to account for differences in organic carbon content, particle density, and resuspension potential between zebra mussel feces versus other sediment material.

    Given the above solids dynamics, the PCB transport and fate follows the conceptual diagram shown in Figure 4b. We assume that PCBS are at local equilibrium with the various solids types, with the partition coefficient based on a chemical-specific Koc and a solid-specific foc. Also, air-water exchange is included as an additional source/sink of PCBS; a standard two-film mass transfer model (DePinto, et al.1994a) is used with a constant gas phase boundary condition. The bioaccumulation and processing of PCBs in the zebra mussels is based on a chemical mass balance in an average individual using the same model framework used for modeling PCB bioaccumulation in Green Bay by Connolly, et al. (1992), in Lake Michigan by Endicott, et al. (1992), in Saginaw Bay by Endicott and Kandt (1994), and in the Buffalo River by DePinto, et al. (1994c). This basic framework represents the latest version of food chain bioaccumulation models that have been evolving over the past 10-15 years (Thomann and Connolly, 1984; Thomann, 1989; Thomann, et al. 1992; Connolly and Pederson, 1988; Gobas, 1993).

    Model Application
    There is an enormous body of literature on PCB transport, fate and bioaccumulation in lakes. However, two important facts suggests that zebra mussels can play a significant role in contaminant cycling and bioaccumulation in the Great Lakes. First, we know that much of the fate of PCBS in lakes is governed by the transport and fate of particulate matter, because PCBS are so hydrophobic. Furthermore, DePinto, et al. (1994b) demonstrated a very close physical coupling between water column and sediment PCBS in Lake Erie through rapid settling and resuspension processes in this system. Second, zebra mussels have been shown to filter out the majority of particulate material from waters in which they have been established, providing direct exposure to particle-bound contaminants. While there have been relatively few reports on zebra mussel bioaccumulation of contaminants, de Kock and Bowmer (1993) demonstrated significant bioaccumulation of cadmium in zebra mussels, with steady state conditions achieved within 40-60 days of exposure. Fisher, et al.(1993; 1994) investigated the uptake and elimination of hydrophobic organic chemicals by zebra mussels, finding relatively high bioconcentration potential. In additional publications from the same overall study, Bruner, et al. (1994a, 1994b) reported that the relative bioaccumulation and assimilation efficiency of two PCB congeners from water, algae, and suspended sediment was strongly dependent on: total PCB concentration and phase distribution, dietary fraction of algae versus seston, and mussel body size and lipid content. They also reported steady-state zebra mussel BAFs (bioaccumulation factors) similar to those found in our screening model and a very significant flux of eliminated PCBS to the benthic environment in the form of feces and pseudofeces.

    In order to examine the impact of zebra mussels on PCB cycling and bioavailability, we configured our screening model to the Lake Erie physical system by dividing the lake into five water column segments (an epilimnion segment for each of the three basins and a hypolimnion segment for the central and eastern basins) with an upper mixed sediment layer below each water column segment where it contacts the sediments. This segmentation is similar to that used by DiToro and Connolly (1980) for their eutrophication modeling in the mid-1970s, except that they also included a metalimnion layer. Hydraulic transport in Lake Erie was based on the previous modeling efforts of DiToro and Connolly (1980) and the recent hydrodynamic modeling in Lake Erie by Schwab and Bedford (Schwab, 1994; Bedford and Schwab, 1994). Solar radiation input and lake temperature regime were obtained from NOAA CoastWatch data, U.S. Weather Bureau data and GLNPO monitoring data. Zebra and quagga mussel spatial densities in Lake Erie for the early 1990s were estimated from data collected in a variety of surveys (e.g., Leach, 1993; Dermot, et al. 1993; Mills, et al. 1993). PCB loading estimates for Lake Erie are available from IJC (GLWQB, 1989) and from other load estimation studies (Strachan and Eisenreich, 1988; Kelly, et al. 1991).

    For the screening analysis that led to the above hypotheses, the model was applied with and without the presence of zebra mussels in the lake but with the same phytoplankton growth kinetics and PCB loading. A constant PCB loading of 1200 Kg/yr was apportioned appropriately to the three surface water column segments (about 2/3 of the total load went to the western basin because of the impact of the Detroit River). A comparison of the steady-state model output for the western basin is presented in Table 1 as average values during the April-October growing season. Note that even though the total water column PCB concentration is decreased in the presence of zebra mussels, the fraction "dissolved" (fd) increases from 0.41 without zebra mussels to 0.55 when zebra mussels are part of the ecosystem. This is because for chemicals like PCBs (with log Koc's in the range of 5-6) the fraction dissolved is a strong function of suspended solids concentration in the range of 1-20 mg/L. In conjunction with the increase in fd, we saw a very significant increase in the mass-specific concentration of PCB in the biotic solids (from 180 to 389 ng/g). Thomann (1989) and others have indicated that food chain bioaccumulation of hydrophobic chemicals are quite sensitive to phytoplankton BCF (bioconcentration factor); therefore, all else being equal zebra mussels may be causing an increased bioaccumulation of PCBS in the pelagic food web.

    Table 1: Comparison of ZM/PCB screening model calculated steady-state response of Lake Erie western basin with and without zebra mussels

    Average value during growing season without zebra with zebra mussels mussels (at 2500/m^2)

    Total suspended solids (mg/L)5.61.9
    Total water column PCB (ng/L)1.10.84
    "Dissolved" PCB (ng/L) 0.450.46
    Total Particulate PCb (ng/L)0.650.38
    Algal PCB content (ng/g dw)180389
    Sediment PCB (ng/g dw)4.88.3
    Mussel PCB (soft tissue) (ng/g ww)--100

    At the same time as we estimated increased biotic solids PCB content in the water column, we found higher sediment levels (from 4.8 to 8.3 ng/g). This is attributed to the higher net deposition rate of solids mediated by the zebra mussel filtering process. Higher sediment PCB concentrations suggests higher PCB bioaccumulation potential through the benthic food chain.

    Conclusion
    The ZEBRA MUSSELS/PCB screening level model described above presents a logical argument for the following hypotheses. Given no change in the PCB loading to Lake Erie, the presence of zebra mussels at reasonable densities will have the following two impacts in the lake:

  • The invasion of zebra mussels in Lake Erie is redirecting the flow of energy, carbon, and phosphorus from the pelagic (grazing) food chain to the benthic (detrital) food chain. Zebra mussels are out competing planktonic herbivores for autochthonously produced organic carbon. By this shunting of particulate matter to the sediments, the zebra mussel invasion is causing an alteration of phytoplankton and phosphorus dynamics in the water column during the growing season and thereby leading to increased total PCB flux to bottom sediments, which creates increased potential for bioaccumulation in the benthic food web.

  • By decreasing the concentration of particulate matter (both biotic and abiotic) in the water column, the zebra mussel invasion is causing an increase in dissolved (bioavailable) and mass-specific algal phase PCB concentrations in the water column, thus also creating a potential for increased bioaccumulation in the pelagic food web.

    Of course, these potential ecosystem implications of the zebra mussel invasion in Lake Erie must be further tested by continued evolution of ecosystem models such as described in this paper. This evolution can only be realized by continued data collection and process experimentation in a range of research areas, including such topics as: developing a quantitative understanding of the fate of particulate matter and associated pollutants that are filtered out of the water column by zebra mussels; and obtaining better estimates of mussel densities, age distribution, and PCB body burdens in the lake as a function of space and time.

    References
    Bedford, K.W. and D.J. Schwab. 1994. The Great Lakes Forcasting System - Lake Erie nowcasts/forecasts. MTS '91, R. Spango (ed.), Marine Tech. Soc., Washington, DC. 432-436.

    Bruner, K.A., S.W. Fisher, and P.F. Landrum. 1994a. The role of the zebra mussel, Dreissena polymorpha, in contaminant cycling: I. The effect of body size and lipid content on the bioconcentration of PCBs and PAHs. J. Great Lakes Res. 20(4):725-734.

    Bruner, K.A., S.W. Fisher, and P.F. Landrum. 1994b. The role of the zebra mussel, Dreissena polymorpha, in contaminant cycling: II. Zebra mussel contaminant accumulation from algae and suspended particles, and transfer to the benthic invertebrate, Gammarus fasciatus. J. Great Lakes Res. 20(4):735-750.

    Chapra, S.C. 1977. Total phosphorus model for the Great Lakes. Jour. Environ. Eng. Div., ASCE. 103(EE2):147-161.

    Connolly, J.P. and C.J. Pederson. 1988. A thermodynamic-based evaluation of organic chemical accumulation in aquatic organisms. Environ. Sci. Technol. 22(1):99-103.

    Connolly, J.P., T.F. Parkerton, J.D. Quadrini, S.T. Taylor, and A.J. Thumann. 1992. Development and application of a model of PCBs in the Green Bay, Lake Michigan walleye and brown trout and their food webs. Technical Report to Large Lakes and Rivers Research Branch, U.S. EPA, ERL-Duluth, Grosse Ile, MI.

    De Kock, W.C. and Bowmer, C.T., 1993. "Bioaccumulation, Biological Effects, and Food Chain Transfer of Contaminants in the Zebra Mussel (Dreissena polymorpha)," In: Zebra Mussels: Biology, Impacts, and Control, Nalepa, T.F. and Schloesser, D.W., eds., Lewis Publishers, Boca Raton, Florida, pp. 503- 533.

    DePinto, J.V., R.K. Raghunathan, P.M. Sierzenga, X. Zhang, V.J. Bierman, Jr., P.W. Rodgers, and T.C. Young. 1994a. Recalibration of GBTOX: An Integrated Exposure Model for Toxic Chemicals in Green Bay, Lake Michigan. Technical Report submitted as part of Cooperative Agreement CR-818560, ERL- Duluth, Large Lakes and Rivers Research Branch, Grosse Ile, MI, 48138, 139 pp.

    DePinto, J.V., P.W. Rodgers, and T. Fiest. "When do sediment-water intereactions control the water column response of large lakesto toxic chemical load reductions?" Oral Presentation at 37th Conference on Great Lakes Research, IAGLR, Windsor, Ontario (June 5-9, 1994b).

    Dermot, R., J. Mitchell, I. Murray, and E. Fear. 1993. Biomass and production of zebra mussels (Dreissena polymorpha) in shallow waters of northeastern Lake Erie. In Nalepa and Schloesser (eds), Zebra mussels: Biology, Impacts, and Control. Lewis Publishers, CRC Press, Inc., Boca Raton, FL. pp. 399- 413.

    Devault, D.S., R. Hesselberg, P.W. Rodgers, and T.J. Feist. 1996. Contaminant trends in lake trout and walleye from the Laurentian Great Lakes. J. Great Lakes Res. 22(4):884-895.

    DiToro, D.M. and J.P. Connolly. 1980. Mathematical Models of Water Quality in Large Lakes, Part 2: Lake Erie. Report No. EPA-600/3-80-065, report to Large Lakes Research Station, ERL-Duluth, Grosse Ile, MI 48138.

    DiToro, D.M., N.A. Thomas, C.E. Herdendorf, R.P Winfield, and J.P. Connolly. 1987. A post audit of a Lake Erie eutrophication model. J. Great Lakes Res. 13(4):801-825.

    Endicott, D.D., W.L. Richardson, and D.J. Kandt. 1992. MICHTOX: A mass balance and bioaccumulation model for toxic chemicals in Lake Michigan. Technical Report, Large Lakes and Rivers Research Branch, U.S. EPA, ERL-Duluth, Grosse Ile, MI.

    Endicott, D.D. and D.J. Kandt. 1994. Far Field Models for Buffalo and Saginaw Rivers and Food Chain Bioaccumulation Model for Saginaw River/Bay. Technical Report prepared for the ARCS/RAM work group. Large Lakes and Rivers Research Branch, U.S. EPA, ERL-Duluth, Grosse Ile, MI.

    Fisher, S.W., Gossiaux, D.C., Bruner, K.A., and Landrum, P.F., 1993. "Investigations of the Toxicokinetics of Hydrophobic Contaminants in the Zebra Mussel," In Zebra Mussels: Biology, Impacts, and Control, Nalepa, T.F. and Schloesser, D.W., eds., Lewis Publishers, Boca Raton, Florida, pp. 469-490.

    Fisher, S.W., Bruner, K.A., and Landrum, P.F., 1994. "Trophic Transfer of Environmental Contaminants by Zebra Mussel, Dreissena polymorpha," Presented at the 37th Conference of the International Association for Great Lakes Research, June 5-9, 1994, Windsor, Ontario.

    Gobas, F.A.P.C. 1993. A model for predicting the bioaccumulation of hydrophobic organic chemicals in aquatic food webs: application to Lake Ontario. Ecological Modeling. 69:1-17.

    Great Lakes Environmental Research Lab/NOAA and U. of Michigan Cooperative Institute for Limnology and Ecosystems Research (CILER). 1994. The Ecological Approach to the Zebra Mussel Infestation in the Great Lakes. Publications Office, NOAA, Ann Arbor, MI.

    Great Lakes Water Quality Board. 1989. Report on Great Lakes Water Quality. International Joint Commission, Windsor, Ontario.

    Griffiths, R.W., Schloesser, D.W., Leach, J.H., and Kovalak, W.P., 1991. "Distribution and Dispersal of the Zebra Mussel (Dreissena polymorpha) in the Great Lakes Region," Canadian Journal of Fisheries and Aquatic Sciences, Vol. 48, pp. 1381-1388.

    Holland, R.E. 1993. Changes in planktonic diatoms and water transparency in Hatchery Bay, Bass Island area, western Lake Erie since the establishment of the zebra mussel. J. Great Lakes Res. 19(3):617-624. Kelly, T.J., J.M. Czuczwa, et al. 1991. Atmospheric and tributary inputs of toxic substances to Lake Erie. J. Great Lakes Res. 17(4):504-516.

    Leach, J.H. 1993. Impacts of the zebra mussel, Dreissena polymorpha, on water quality and spawning reefs in western Lake Erie. In Nalepa and Schloesser (eds), Zebra mussels: Biology, Impacts, and Control.Lewis Publishers, CRC Press, Inc., Boca Raton, FL. pp.381-398.

    Makarewicz, J.C. 1993. Phytoplankton biomass and its species composition in Lake Erie, 1970 to 1987. J. Great Lakes Res. 19(2):258-274.

    Makarewicz, J.C. and P.E. Bertram (Eds.). 1993. Evidence for the Restoration of Lake Erie. Special Issue of J. Great Lakes Res. 19(2):197-309.

    Mills, E.L., R.M. Dermott, E.F. Roseman, D. Dustin, E. Melina, D.B. Conn, and A.P. Spidle. 1993. Colonization, ecology, and population structure of the "quagga" mussel (Bivalvia: Dressenidae) in the lower Great Lakes. Can. J. Fish. Aquat. Sci. 50:2305-2314.

    Nicholls, K.H. and Hopkins, G.J., 1993. "Recent Changes in Lake Erie (North Shore) Phytoplankton: Cumulative Impacts of Phosphorus Loading Reductions and the Zebra Mussel Introduction," Journal of Great Lakes Research, Vol. 19, No. 4, pp. 637-647.

    Raghunathan, R.K. 1990. Development of a dynamic mass balance model for PCBs in Green Bay. M.S. Thesis, Clarkson University, Potsdam, NY.

    Rodgers, P.W., J.V. DePinto, T. Slawecki, and W. Booty. 1987. LTI Toxics Model Application: PCBs in Lake Ontario - An Exploratory Analysis. Report to IJC Task Force on Chemical Loadings, Windsor, Ontario.

    Schwab, D.J. 1994. Initial implementation of the Great Lakes Forecasting System: a real-time system for predicting lake circulation and thermal structure. Water Poll. Res. J. Can. 2/3:302-220.

    Schneider, D.W., 1992. "A Bioenergetics Model of Zebra Mussel, Dreissena polymorpha, Growth in the Great Lakes," Canadian Journal of Fisheries and Aquatic Sciences, Vol. 49, pp. 1406-1416.

    Spidle, A.P., J.E. Marsden, and B. May. 1994. Identification of the Great Lakes quagga mussel as Dreissena bugensis from the Dneiper River, Ukraine on the basis of allozyme variation. Can. J. Fish. Aquat. Sci.51:1485-1489.

    Strachan, W.M.J. and S.J. Eisenreich. 1988. Mass Balancing of Toxic chemicals in the Great Lakes: The Role of Atmospheric Deposition. Report to Science Advisory Board, IJC, Windsor, Ontario.

    Thomann, R.V. and D.M. DiToro. 1983. Physicochemical model of toxic substances in the Great Lakes. J.Great Lakes Res. 9(4):474-496.

    Thomann, R.V. and J.P. Connolly. 1984. Model of PCB in the Lake Michigan lake trout food chain. Environ. Sci. Technol. 18(2):65-71.

    Thomann, R.V. 1989. Bioaccumulation model of organic chemical distribution in aquatic food chains. Environ. Sci. Technol. 23(6):699-707.

    Thomann, R.V., J.P. Connolly and T.F. Parkerton. 1992. An equilibrium model of organic chemical accumlation in aquatic food webs with sediment interaction. Environ. Toxicol. Chem. 11:615-629

    . Vollenweider, R.A. 1975. Input-output models with special reference to the phosphorus loading concept in limnology. Schweizerische Zeitschrift fur Hydrologie. 37:53-84.

    Zebra Mussels and Benthic Macroinvertebrate Communities of Southwestern Lake Ontario and Selected Tributaries: Unexpected Results?
    James M. Haynes
    Center for Applied Aquatic Science and Aquaculture
    Department of Biological Sciences
    State University of New York College at Brockport
    350 New Campus Drive
    Brockport, NY 14420-2973

    Abstract
    Since the zebra mussel (Dreissena polymorpha) colonized the Great Lakes ecosystem, ecologists and managers have expressed concern about potential impacts it would have on native benthic macroinvertebrate communities in lakes and streams. We compared post-Dreissena invasion data (1991-1992) with pre-invasion data (1983) from the same sites in Lake Ontario. Dreissena was the overwhelmingly dominant taxon in 1991-1992. Nevertheless, the overall abundance of other benthic macroinvertebrates, and the number of taxa collected, were greater following establishment of Dreissena. Our study failed to provide evidence that Dreissena has induced a population decline in any non-bivalve taxon that was present in 1983. The invasion of Dreissena and other recent environmental changes appear to have created conditions more favorable for most benthic macroinvertebrate taxa in the nearshore region of Lake Ontario. We also asked why zebra mussels have not colonized many creeks in western New York, some with apparently ideal habitat, that are fed partially by water from the Erie Canal which is a source of Dreissena larvae. Counts of Dreissena larvae and concentrations of chlorophyll a were much higher in the canal than in our study creek, yet water quality, current velocity, and particulate organic carbon concentrations in the creek and the canal were very similar. Four factors appear to limit colonization of creeks by the zebra mussel: 1) Retention of larvae by wetlands through which discharges from the canal often flow, 2) Filtering of phytoplankton and larvae by dense beds of adult zebra mussels often found at the beginning of channels connecting the canal to creeks, 3) Inappropriate food quality (e.g., lack of small-diameter phytoplankton with important fatty acid constituents) reaching creeks from the canal, or 4) Muddy substrates inappropriate for attachment and filter feeding by Dreissena.

    Introduction
    First reported in the Great Lakes in 1988, the zebra mussel (Dreissena polymorpha) has colonized many portions of the Great Lakes basin and connected watersheds. The spread of Dreissena has generated concern among ecologists for several reasons, particularly their ability to: 1) Completely cover and change the physical structure of hard substrates and 2) Reduce open lake phytoplankton biomass by filter feeding which likely impacts pelagic food webs dependent upon microscopic algae (e.g., zooplankton, alewives, stocked salmon). Since 1990, researchers at State University of New York at Brockport have studied the effects of zebra mussels on benthic macroinvertebrate communities associated with bottom substrates in southwestern Lake Ontario and in several western New York creeks connected to the Erie Canal and Lake Ontario.

    Studies in Southwestern Lake Ontario
    Because Dreissena and other benthic macroinvertebrate taxa are fairly immobile, often occupy the same substrates, and in some instances may consume similar foods, changes in previously established benthic macroinvertebrate communities following Dreissena invasion are possible. However, scarcity of pre-invasion data has made assessment of Dreissena impacts on pre-established macroinvertebrate taxa difficult. In 1991-1992, we used the same location, study design and sampling methods as those of Bader (1985), who quantified abundances of benthic macroinvertebrate taxa in 1983, seven years before the establishment of Dreissena in Lake Ontario. Full details are in Stewart and Haynes (1994) and Stewart (1993).

    In 1991-1992, zebra mussels comprised up to 93% of the macroinvertebrates (insects, crustaceans, worms, snails, etc.) collected, replacing the amphipod sideswimmer (Gammarus fasciatus) which was the numerically dominant taxon in 1983. However, the total abundance of non-Dreissena macroinvertebrates was significantly greater in 1991-1992 (range: 1,316 + 170 to 5,267 + 523 /m2) than in 1983 (range: 127 + 41 to 1,159 + 107 /m2). Taxa showing the greatest increases in abundance included the annelid worms (Manayunkia speciosa, Spirosperma ferox) and unidentified tubificids; the gastropod snails (Helisoma anceps, Physa heterostropha, Stagnicola catascopium, Valvata tricarinata, Goniobasis livescens, Amnicola limosa); the sideswimmer (Gammarus fasciatus); the trichopteran insect (Polycentropus sp.); and the decapod crayfish (Orconectes propinquis). No taxon was less abundant in 1991-1992 than in 1983, and comparisons of macroinvertebrate community similarity in 1983 and 1991-1992 (range: 50 to 99% similarity) indicated that previously established taxa did not change substantially between 1983 and 1991-1992. The number of taxa collected was significantly greater in 1991-1992 (maximum: 32 per sample) than in 1983 (maximum: 15 per sample). While taxa of numerical importance in 1983 remained important in 1991-1992, some taxa of little numerical importance in 1983, such as the oligochaete worms (Stylaria lacustris, Potamothrix vejdovskyi, S. ferox), and the snail (Amnicola limosa), were of increased importance in 1991-1992. No taxon exhibited a significant population decline between 1983 and 1991-1992.

    Population changes of some taxa are similar to those reported by other researchers who have studied impacts of Dreissena on benthic macroinvertebrate communities in Lakes Erie and St. Clair (Dermott et al. 1993, Griffiths 1993). Although other factors may have contributed to the changes observed, our results support theories that Dreissena is facilitating the transfer of energy to bottom organisms by pseudofecal/fecal deposition, that mussel colonies are providing habitat for additional invertebrate taxa, and that predicted disasterous changes in native benthic macroinvertebrate communities (except for freshwater clams) have yet to occur in Lake Ontario.

    Dreissena may facilitate transfer of nutrients to benthic macroinvertebrates by filter-feeding on open water phytoplankton and subsequently depositing wastes on the bottom. Dreissena feces and pseudofeces are important in diets of benthic macroinvertebrates that eat detritus (dead organic material). Facultative or obligatory detritivores with population increases in our study were S. ferox and other tubificid worms, the snails (P. heterostropha, V. tricarinata), the amphipod (G. fasciatus), and the insect (Polycentropus sp.). Other researchers have found the abundance and biomass of many benthic invertebrates, including detritivorous oligochaete worms and midgefly (chironomid insect) larvae, to be greatest among clumps of Dreissena where feces and pseudofeces accumulate.

    Enhanced substrate complexity also may contribute to increased abundance and diversity of macroinvertebrates. By creating an interstitial network that may increase refugia available to other benthic organisms, Dermott et al. (1993) suggested that Dreissena was responsible for increases in Gammarus sp. observed on bedrock substrates colonized by Dreissena. Griffiths (1993) likewise attributed population increases of leeches, snails, sideswimmers, (Polycentropus) and the midgefly (Polypedilum sp.) in Lake St. Clair to increased substrate heterogeneity provided by Dreissena. Of these taxa, only leeches and Polypedilum failed to show significant population increases at our study sites between 1983 and 1991-1992.

    Dreissena may indirectly create benthic habitat as well. Filter-feeding improves water clarity through removal of suspended particles. The resulting increase in the lighted water zone, in combination with increased transfer of nutrients to the lake bottom by Dreissena, may promote growth of benthic vascular plants (aquatic weeds). Positive relationships between benthic algae and populations of nematode round worms, naidid oligochaete worms, leeches, snails (Gyraulus sp., Helisoma sp., Physa sp., Valvata sp., Goniobasis sp., Amnicola sp.), sideswimmers, mayfly (ephemeropteran insect) larvae and midgefly larvae have been reported from the Great Lakes in the past (Cook and Johnson 1974, Barton and Hynes 1978). Griffiths (1993) believed that increased densities of submerged vascular plants and benthic algae following colonization of Lake St. Clair by Dreissena contributed to the observed increase in macroinvertebrate populations there. A variety of filamentous green algae were present at our study sites in 1983 and in 1991-1992, and macroinvertebrates, especially the sideswimmer (G. fasciatus), were associated with the algae.

    While most macroinvertebrate population changes observed in our study may be attributable to Dreissena, these changes may also reflect changes in water quality or habitat conditions that are unrelated to the Dreissena invasion. Phosphorus abatement programs have contributed to declines in total phosphorus concentrations throughout Lake Ontario since the mid-1970s, but assessing the effects that phosphorus abatement has had and will continue to have on benthic macroinvertebrate populations is problematic. Johnson and MacNeil (1986) attributed declines in the abundances of some oligochaete worm, sphaeriid clam and isopod/scud taxa in the Bay of Quinte to reductions in phosphorus loading to Lake Ontario. Barton (1986) observed declines in total benthic macroinvertebrate abundance in areas undergoing rapid deeutrophication, but noted that species diversity often increased under such conditions. Increased overall abundance of benthic macroinvertebrates, including at least one worm taxon (S. ferox) known to inhabit nutrient-rich habitats, suggest nutrient deposition by Dreissena has more than compensated for oligotrophication processes in Lake Ontario between 1983 and 1991-1992. Increased water clarity resulting from declining phosphorus concentrations, in combination with Dreissena filter-feeding, also may lead to a deeper and warmer epilimnion (Mazumder 1990) which may further increase benthic production.

    Other studies conducted within the Great Lakes strongly suggest Dreissena is threatening clams of the family Unionidae by settling on their shells and possibly inhibiting their ability to feed, respire, and reproduce (cf. Mackie 1991). Our study did not focus on freshwater clams and failed to show that Dreissena is negatively affecting other benthic macroinvertebrate taxa. Our data suggest that Dreissena has thus far had a positive impact on many benthic macroinvertebrate species in southwestern Lake Ontario.

    Studies in Western New York Creeks
    The New York State Erie Barge Canal, with a direct connection to Lake Erie, was colonized by zebra mussels in 1989. Many of the creeks and rivers crossed by the canal receive water from it from April through November. Dense colonies of zebra mussels were observed at canal water outfalls to these creeks as early as 1990, but none of six creeks examined in Monroe County, New York has zebra mussels more than 100 m downstream from where canal discharges enter.

    Salmon Creek was chosen for study (see Miller and Haynes 1997 and Miller 1994 for full details of methods and results) because there was no apparent reason why zebra mussels should not successfully colonize the creek. Rocky substrates suitable for attachment are abundant, and invertebrate and fish communities indicate a reasonably healthy environment. Aside from agricultural and suburban run-off common to all streams in the region, there is no evidence of point source pollution in the watershed. We examined water quality (temperature, pH, calcium carbonate concentration), physical conditions (current, substrate), and biological conditions (food supply, predation) to learn why zebra mussels have not become established.

    Water temperature, pH, calcium carbonate concentration, and current velocity in the Erie Canal and Salmon Creek did not differ significantly during the sampling period (Table 1), and they can be eliminated as factors reducing zebra mussel abundance in the creek. In fact, physical and chemical conditions in the creek were generally in the optimal ranges for the survival, growth and reproduction of zebra mussels (Table 2).

    Table 1. Comparison of mean physical, chemical and biological conditions most relevant to zebra mussel biology in the Erie Canal and Salmon Creek

    Parameter Erie Canal Salmon Creek Significant
    Temperature (oC) 19.6 19.4 No
    ph 7.9 8.0 No
    Calcium (mg/L) 62.9 72.5 No
    POC (mg/L) 21.5 21.6 No
    Veligers (#/L) 14.3 0.2 Yes
    Chlorophyll a (mg/L) 7.9 4.6 Yes
    Velocity (m/sec) 0.09 0.05 No
    Discharge (m3/sec) 0.07 0.08 No

    Table 2. Tolerated and optimal ranges of important environmental parameters for zebra mussels. Makarewicz reported water quality data for nearby Larkin, Buttonwood and Northrup Creeks.

    Parameter Tolerated Range Optimum Range Makarewicz (1989) Salmon Creek
    Temperature (oC) 0 - 30 a 20 - 25 a 1- 27 9 - 27
    ph 4.6 - 9.5 b 8.2 - 8.6 b 7.5 - 8.5 7.3 - 8.3
    Calcium (mg/L) > 40 a,c > 50 a,c 61 - 68 49 - 118
    Current (m/sec) > 0 - 2.5 d,e 0.1 - 1 d,e < 1 0.03 - 0.16
    Food Particle Size (m) 1 - 450 b,f,g 15 - 45 b,f,g not reported not reported

    a Kovalak 1989, b Morton 1971; c Sprung 1993, d Smirnova and Vinogradov 1990, e O'Neill and MacNeill 1991, f Sprung and Rose 1988, g Ten Winkel and Davids 1982

    Fish of 11 species were collected from Salmon Creek and the canal outfall channel, but only one zebra mussel was found in the stomach of one fish. Although crayfish abundance in Lake Ontario increased after colonization by the zebra mussel (Stewart 1993), and predation on zebra mussels by crayfish has been observed in controlled settings, the abundance of crayfish in Salmon Creek does not appear to have changed. Predation also does not appear to be a likely factor limiting zebra mussel abundance in Salmon Creek.

    Initially, we hypothesized that food supplies for zebra mussels (measured as particulate organic carbon, POC, in the water) would decrease as canal water moved over the existing zebra mussel colony in the outfall channel and was diluted by Salmon Creek, but POC did not differ between the canal and the creek (Table 1). Thus, food quantity, as measured by POC, was eliminated as a factor limiting zebra mussel colonization of Salmon Creek.

    The abundance of veligers and the concentration of chlorophyll a dropped sharply after water left the Erie Canal (Table 1). In fact, veliger counts dropped 50 percent or more and chlorophyll a levels dropped by an average of 87% less than 15 m down the channel leading from the canal to the creek. Maximum densities of Dreissena larvae were 55/L in the canal and 2.3/L in the creek. Maximum chlorphyll a levels were 21.7 mg/L in the canal and 7.3 mg/L in the creek. What reduced veliger counts and chlorophyll a concentration after the canal outfall?

    Near the base of the canal outfall channel to Salmon Creek is a dense colony of adult zebra mussels and a small wetland. Part of the water flowing out of the canal forms a back eddy that flows through the wetland. Adult zebra mussels were observed attached to vegetation in the wetland, suggesting that some veligers become trapped and settle in the wetland. That this may occur in Salmon Creek was supported by sampling in Brockport Creek. The canal discharge to Brockport Creek meanders through a wetland before reaching the creek. No adult zebra mussels were found in the creek, but chlorophyll a levels were much higher after water passed through the wetland than they were in the canal. This suggests that the wetlands associated with canal outfall channels produce phytoplankton but prevent veligers from reaching the creeks.

    Literature published before 1993 indicated that food particle size, not composition, is the critical aspect of diet for zebra mussels; 15 to 45 m is the preferred size range, but 1 to 450 m particles can be filtered and ingested (Ten Winkel and Davids 1982). Later studies suggest that food quality also is important. Stoeckman and Gerton (Ohio State University, Columbus, Ohio, personal communication) found better survival among cultured zebra mussels using a commercial diet of marine algae higher in fatty acids than the control diet. Vanderploeg et al. (1996) reported that the key to raising Dreissena in culture is providing the right algae, particularly freshwater Chlorella minutissima and marine Rhodomonas minuta, both about 3 m in diameter and rich in long chain polyunsaturated fatty acids. However, Wright et al. (1996) reported that two Dreissena species survived and grew better on small phytoplankton high in saturated fatty acids. Clearly, small food size is important to veligers, but more work is needed to precisely define key components of the Dreissena diet and how diet requirements may influence where mussels can colonize successfully.

    The Erie Canal has many of the physical and biological properties of a lake, among which is the presence of phytoplankton and bacteria suitable for zebra mussel feeding. Chlorophyll a levels in the canal were consistently higher than levels in Salmon Creek. The high abundance of zebra mussels in the canal is undoubtedly related to its rich food supply. Because phytoplankton do not readily occur in streams and 87% of canal phytoplankton (measured as chlorophyll a) is filtered by adult Dreissena before reaching the creek, it is quite likely that the POC of Salmon Creek does not meet the qualitative nutritional requirements of zebra mussels.

    Just as adult zebra mussels at the base of the canal discharge to Salmon Creek appear to filter most phytoplankon out of the water, it is likely they are also feeding on the larvae (Smirnova and Vinogradov 1990) coming from the canal and create the 60% reduction in numbers observed.. The dense population of adult Dreissena in the discharge channel appears to be a "biotic sponge" which removes veligers and phytoplankton from the canal water flowing into Salmon Creek, thus depriving the creek of appropriate quality food (phytoplankton) and a source of larvae to support colonization of the creek.

    Why have zebra mussels not colonized other creeks in the region? The wetland between the canal and Brockport Creek apparently prevents veligers from reaching the creek. Near their outfalls, Allens and Northup Creeks have predominately muddy bottoms unsuitable for zebra mussel attachment and filter feeding. Thus, it appears that zebra mussels are having a difficult time colonizing streams fed by the Erie Canal in Monroe County, New York.

    Conclusion
    By occurring in large clumps and filter-feeding intensively, Dreissena may benefit other macroinvertebrate taxa in Lake Ontario by increasing the complexity of benthic substrate and by increasing the flow of energy to benthic environments. While some factors (e.g., increased water clarity due to Dreissena biofiltration and phosphorus abatement programs) may be acting synergistically to favorably impact benthic organisms, biodeposition by Dreissena may keep benthic food resources at high levels despite overall declines in fertility of the Lake Ontario ecosystem. Given the appropriate physical habitat and water quality conditions and an abundant source of veligers, four factors appear to limit colonization of regional creeks by zebra mussels: 1) Partial retention of zebra mussel veligers by wetlands through which the canal discharges often flow, 2) Filtering of phytoplankton and veliger larvae by adult Dreissena often found at the beginning of outfall channels from the canal to creeks, 3) Inappropriate food quality (e.g., lack of small-diameter phytoplankton with important fatty acid constituents) reaching creeks from the canal, or 4) Muddy substrates inappropriate for attachment and filter feeding by Dreissena. Considering the many dire warnings of impending ecological disaster after invasion by zebra mussels, these results are encouraging. Our research team continues to monitor for potential longer term changes, but so far it appears that the zebra mussel has not greatly changed benthic communities in lakes and streams in our part of the Great Lakes basin.

    Acknowledgments
    I thank Timothy W. Stewart and Steven J. Miller who wrote the M.S. theses upon which this paper is based.

    Literature Cited
    Bader, A.P. 1985. Dynamics of benthic macroinvertebrates inhabiting an artificial reef and surrounding areas in southwestern Lake Ontario. M.S. thesis, SUNY College of Environmental Science and Forestry, Syracuse, NY.

    Barton, D.R. 1986. Nearshore benthic invertebrates of the Ontario waters of Lake Ontario. J. Great Lakes Res. 12:270-280.

    Barton, D.R., and Hynes, H.B.N. 1978. Wave-zone macrobenthos of the exposed Canadian shores of the St. Lawrence Great Lakes. J. Great Lakes Res. 4:27-45.

    Cook, D.G., and Johnson, M.C. 1974. Benthic invertebrates of the St. Lawrence-Great Lakes. J. Fish. Res. Board Can. 31:763-782.

    Dermott, R., Mitchell, J., Murray, I., and Fear, E. 1993. Biomass and production of zebra mussels (Dreissena polymorpha) in shallow waters of northeastern Lake Erie. In: Zebra Mussels: Biology, Impacts, and Control, T.F. Nalepa and D.W. Schloesser (Eds.), pp. 399-413. Lewis Publishers, Boca Raton, FL.

    Griffiths, R.W. 1993. Effects of zebra mussels (Dreissena polymorpha) on the benthic fauna of Lake St. Clair. In: Zebra Mussels: Biology, Impacts, and Control, T.F. Nalepa and D.W. Schloesser (Eds.), pp. 415-437. Lewis Publishers, Boca Raton, FL.

    Johnson, M.G. and McNeil, O.C. 1986. Changes in abundance and species composition in benthic macroinvertebrate communities of the Bay of Quinte, 1966-1984. In: Project Quinte: point-source phosphorous control and ecosystem response in the Bay of Quinte, Lake Ontario, C.K. Minns, D.A. Hurley and K.H. Nicholls (Eds.), pp. 177-189. Can. Spec. Publ. Fish. Aquat. Sci. 86.

    Kovalak, W.P. 1989. Life history and biology of the zebra mussel (Dreissena polymorpha).Engineering Research Report 89A90-3. 7 p. (Obtained from NY Sea Grant Zebra Mussel Clearing House, SUNY College at Brockport, NY).

    Mackie, G.L. 1991. Biology of the exotic zebra mussel, Dreissena polymorpha, in relation to native bivalves and its potential impact in Lake St. Clair. Hydrobiologia 219:251-268.

    Makarewicz, J.C. 1989. Chemical analysis of water from Buttonwood, Larkin, Round Pond and Northrup Creeks, Lake Ontario basin west, May 1987-May 1988. Monroe County Department of Health. Rochester, NY.

    Mazumder, A. 1990. Ripple effects: how lake dwellers control the temperature and clarity of their environment. New York, NY, The Sciences, New York Academy of Sciences, Nov./Dec. pp. 39-42.

    Miller, S.J. 1994. An analysis of factors potentially limiting the abundance of the zebra mussel (Dreissena polymorpha) in Salmon Creek, Monroe County, New York. M.S. thesis. SUNY College at Brockport. Brockport, NY.

    Miller, S.J., and Haynes, J.M. 1997. Factors limiting colonization of western New York creeks by the zebra mussel (Dreissena polymorpha). J. Freshwater Ecol. In press.

    Morton, B. 1971. Studies on the biology of Dreissena polymorpha Pall. V. Some aspects of filter feeding and the effect of micro-organisms upon the rate of filtration. Proc. Malacological Soc. London 39:289-301.

    O'Neill, C.R., Jr., and MacNeill, D.B. 1991. The zebra mussel (Dreissena polymorpha): an unwelcome North American invader. Coastal Resource Fact Sheet. NY Sea Grant. SUNY College at Brockport. Brockport, NY.

    Smirnova, N.F., and Vinogradov, G.A. 1990. Biology and ecology of Dreissena polymorpha from the European USSR. Presented at the workshop on Introduced Species in the Great Lakes: Ecology and Managment. Saginaw, MI. 26-28 Sept.

    Sprung, M. 1993. The other life: an account of present knowledge of the larval phase of Dreissena polymorpha. pp. 39-53. In: Zebra Mussels: Biology, Impacts, and Control, T.F. Nalepa and D.W. Schloesser (Eds.). Lewis Publishers, Boca Raton, FL.

    Sprung, M., and Rose, U. 1988. Influence of food size and food quantity of the feeding of the mussel Dreissena polymorpha. Oecologia 77:526-632.

    Stewart, T.W. 1993. Benthic macroinvertebrate community changes following zebra mussel colonization of southwestern Lake Ontario. M.S. thesis, SUNY College at Brockport, Brockport, NY.

    Stewart, T.W., and Haynes, J.M. 1994. Benthic macroinvertebrate communities of southwestern Lake Ontario following invasion of Dreissena. J. Great Lakes Res. 19(2):479-493.

    Ten Winkel, M. E. H., and Davids, C. 1982. Food selection by Dreissena polymorpha Pallas (Mollusca: Bivalvia). Freshwater Biology 12:553-558.

    Vanderploeg, H.A., Liebig, J.R., and Cook, A.A. 1996. Evaluation of different phytoplankton for supporting development of zebra mussel larvae (Dreissena polymorpha): the importance of size and polyunsaturated fatty acid content. J. Great Lakes Res. 22:36-45.

    Wright, D.A., Setzler-Hamilton, E.M., Magee, J.A., and Harvey, H.R. 1996. Laboratory culture of zebra (Dreissena polymorpha) and quagga (D. bugensis) mussel larvae using estuarine algae. J. Great Lakes Res. 22:46-54.

    Surveillance for Ruffe in the Great Lakes - An Overview
    Sandra M. Keppner
    U.S. Fish and Wildlife Service
    Lower Great Lakes Fishery Resources Office
    405 North French Road
    Amherst, NY 14228

    Thomas R. Busiahn
    Ashland Fishery Resources Office
    U.S. Fish and Wildlife Service
    2800 Lake Shore Drive East
    Ashland, WI 54806

    Jerry McClain
    Alpena Fishery Resources Office
    U.S. Fish and Wildlife Service
    145 Water Street
    Alpena, MI 49707

    Gord Johnson
    Ontario Ministry of Natural Resources
    Lake Superior Management Unit
    435 South James Street
    Thunder Bay, Ontario P7E 6S8

    Abstract
    The establishment of nonindigenous aquatic species is a perturbation that challenges resource managers in the Great Lakes basin to understand the dynamics of the present system, and predict patterns for the future. The effects of invasive species and their eventual distribution are difficult to predict. Surveillance programs are used by managers as a means of identifying newly established populations, tracking range expansions, assessing the impacts of introduced species on native populations, and evaluating control or management strategies.

    The ruffe, Gymnocephalus cernuus, like the zebra mussel, is believed to have been introduced to the Great Lakes through ballast water discharge. Since its discovery, surveillance programs have continued to track range expansion and monitor species abundance in areas where ruffe have already colonized as well as in areas where ruffe are likely to colonize. In addition to conducting field surveys, the education of public water users has been instrumental to increasing surveillance efforts. In several cases, recreational anglers were first to report the presence of ruffe at new locations, findings that were later verified with fish sampling gear. Communication with the maritime industry has guided surveillance efforts to shipping ports where ruffe may colonize via ballast water transport. The probability of detection by surveillance sampling of new colonies of ruffe appears to be rather high, because ruffe are highly selective toward favorable habitat at low population densities.

    Western Lake Superior surveys have documented the natural movement of ruffe, changes in fish species abundance, and the ability of ruffe to become the dominant fish species. Preliminary results suggest declines in native fish populations following the successful establishment of ruffe. Surveys in areas where ruffe have not yet invaded provide information regarding native or currently existing populations, without which it is difficult to determine the effects of an invasion. The proactive management approach exemplified by the Great Lakes Ruffe Surveillance Program, conducted by the U.S. Fish and Wildlife Service and the Ontario Ministry of Natural Resources, provides a model response to species invasions.

    Introducation
    The introduction of nonindigenous species is a perturbation that alters ecosystem dynamics, affecting population and community structure as well as species abundance and diversity. Understanding the ecological effects of introduced nonindigenous species, including costs and benefits, challenges resource managers. In the Great Lakes, over 140 aquatic organisms have been introduced since the 1800s (Dermott 1997, Mills et al. 1993, Zaranko 1996). The rate of introductions has increased as mechanisms of travel and transport advance and waterways are connected through man-made canals and channels. For example, the opening of the St. Lawrence Seaway coincides with increasing numbers of introductions (Mills et al. 1993).

    Upon the discovery of a newly introduced species, managers are immediately challenged to determine the extent of the invasion, including the current distribution, abundance, and immediate impacts of the organism. In addition, managers are asked to assess or predict range expansions, pathways of dispersal, and potential long-term impacts to invaded waters. These factors contribute to the decision-making process regarding the need for control as well as the potential success of control alternatives.

    To make these assessments and predictions, it seems clear that managers need information regarding the ecological conditions prior to an invasion, including species composition and abundance, interactions between species, and habitat availability. Further, managers need to gain an understanding of the biology and ecology of the introduced species and its potential mechanisms of dispersal. Information regarding the new organism's role in its native waters or other areas of introduction contributes significantly to the decision-making process. However, pre-invasion information including investigations of the community as a whole (i.e. all trophic levels) is often sparse or incomplete.

    The need for surveillance programs is emphasized by the introduction of nonindigenous species. Surveillance programs provide managers with the opportunity to 1) identify newly established populations early, 2) track or detect range expansions, 3) estimate potential impacts of introductions or range expansions by gathering baseline data on pre-existing populations and habitat, and 4) evaluate control or management strategies. The information that can be obtained through effective surveillance programs may contribute significantly to the decision-making process regarding the risk of further expansion and impacts to native fauna, as well as the need for control, or the continuation of control implementation.

    The Great Lakes Ruffe Surveillance Program (the Program) exemplifies how such a program can be conducted effectively. The Program provides a model illustrating how management agencies can respond effectively to surveillance needs for nonindigenous species introductions. The Program has played and continues to play a critical role in management decisions.

    Background
    The ruffe (Gymnocephalus cernuus) is a small percid fish thought to have been introduced into the Great Lakes in the early to mid 1980s via ballast water discharge from trans-Atlantic shipping. It was first collected and identified in 1987 in fish collections from the St. Louis River, the westernmost tributary of Lake Superior, by Wisconsin Department of Natural Resources (Pratt et al. 1992). Re-examination of ichthyoplankton samples collected in 1986 confirmed their earliest known presence in the river (Simon and Vondruska 1991, Pratt et al. 1992). Since then, ruffe populations have continued to increase in abundance and distribution, expanding their range along the southwestern shoreline of Lake Superior. In 1991, ruffe were collected in Thunder Bay, Ontario, likely transported there in the ballast water of ships departing the St. Louis River (Busiahn, 1997). By 1993, ruffe had expanded as far east as the Bad River, Wisconsin, approximately 156 km (97 mi) from the St. Louis River (Slade et al. 1994). In 1994, ruffe continued their eastward expansion, being collected in the Ontonagon River, Michigan, approximately 300 km (186 mi) east of the St. Louis River (Slade et al. 1995). In 1995, ruffe were collected for the first time outside of Lake Superior at the mouth of the Thunder Bay River, Lake Huron near Alpena, Michigan (Kindt et al. 1996) (Figure 1). This expansion marked a critical turning point in the management or control of ruffe in the Great Lakes.

    Figure 1: Locations where the presence of ruffe has been confirmed

    The importance of and need for consistent surveillance was recognized early in the history of the ruffe invasion. In 1992, a Ruffe Control Committee (the Committee) was convened by the national Aquatic Nuisance Species Task Force. The Committee was charged to develop and refine a control program in accordance with the Nonindigenous Aquatic Nuisance Prevention and Control Act of 1990 (P.L.101-636) that minimizes the risk of harm to the environment and to the public health and welfare (Busiahn and McClain 1995). In developing their plan for control, the importance of surveillance became more apparent as a means of early detection to design and target areas for control as well as to evaluate proposed control efforts.

    Surveillance was identified as one of the eight primary objectives of the Ruffe Control Program. The objective specifically required a coordinated program that would successfully identify newly established populations of ruffe. This was distinguished within the Ruffe Control Program from efforts to investigate established populations of ruffe and the subsequent changes in ecosystem dynamics following their successful establishment. The Ruffe Control Program also identifies the U.S. Fish and Wildlife Service (the Service) and the Ontario Ministry of Natural Resources (OMNR) as the lead agencies coordinating surveillance efforts.

    Ruffe Surveillance Program
    The Ashland Fishery Resources Office is the main contact for all ruffe activities conducted by the Service and has led surveillance efforts in the upper Great Lakes, including Lakes Superior, Michigan, and Huron, since 1992. The collection of ruffe near Alpena, Michigan, in Lake Huron in 1995 initiated the involvement of the Service's Alpena Fishery Resources Office. Since 1996, surveys in Lake Huron and northern Lake Michigan have been conducted by the Alpena Fisheries Resources Office. Surveillance efforts on the lower Great Lakes were initiated in 1993 by the Service's Lower Great Lakes Fishery Resources Office. And finally, Ontario's program, initiated in 1991, is led by the Lake Superior Management Unit in Thunder Bay, Ontario.

    The current Program implemented by both agencies consists of three components: field surveys, a mail survey, and public education. This report will provide a brief overview of each component and how they have contributed to the overall success of the Program.

    Field Surveys
    The primary objective of the field survey component is to locate new populations of ruffe and describe their age/size composition. The second objective is to describe the fish community at each location surveyed. The field survey component, if conducted effectively, can provide managers with information towards early detection, range expansion, pre-existing populations, and control evaluation.

    Field surveys for ruffe throughout the Great Lakes target estuaries, bays, river mouths, and waterways on the periphery of the ruffe's range and in or near shipping ports where ruffe could potentially colonize through inter- or intra-lake shipping activities. Most of the sampling targets the deepest habitat available at any given site, including natural holes or channels in rivers and estuaries or in the dredged shipping channels. The primary means of fish collection is bottom trawling; however, electrofishing, seines, gill-nets, fyke nets, and modified windemere traps are also used. Each survey site is visited two to three times between May and October based on habitat suitability and distance from the periphery of the known range.

    Since field surveys were initiated in 1992 by the Ashland Fishery Resources Office, the number and location of survey sites has changed in response to expanding ruffe populations. Figure 2 represents a comprehensive overview of all the sites surveyed in the upper Great Lakes, 1992 - 1996. Surveillance at all of these sites, with one exception, was being conducted by the Ashland Fishery Resources Office. The one exception is the site in Thunder Bay, Ontario. Trawling effort there is led by the OMNR. As mentioned previously, the Alpena Fishery Resources Office is now participating in the Program, leading surveillance efforts in Lake Huron and northern Lake Michigan. Surveillance in Lakes Erie and Ontario is conducted by the Lower Great Lakes Fishery Resources Office. Seven sites on Lake Erie are surveyed, and one site on Lake Ontario (Figure 3). Surveys in the lower lakes target areas where ruffe may potentially colonize through inter-lake ballast transfers.

    Figure 2: Locations Surveyed for Ruffe in the Upper Great Lakes


    Sampling procedures include measuring environmental and water quality parameters, such as water temperature, dissolved oxygen levels, depth, and Secchi transparency, prior to each trawl. The average depth of the trawled area, average tow speed, and total time is recorded for each tow. All fish collected are identified, counted, and measured to the nearest mm (TL). When more than 50 individuals of any single species are collected in a single tow, a random selection of 50 individuals are measured, providing an adequate sample size to infer size composition of the total number. Fish that can not be identified in the field, such as small juveniles, are preserved and identified in the lab. Any ruffe that are collected are measured, placed in zip-lock bags and frozen for later analysis. Age determination of ruffe is determined in the laboratory by examining the second dorsal spine.

    With approximately 50 sites surveyed annually, reporting and analyzing results can become exhaustive. An annual report, summarizing the expansion of ruffe confirmed that year, as well as common species collected, has been prepared and distributed since 1992 (Slade and Kindt, 1992; Slade et al. 1994; Slade et al. 1995; Kindt et al. 1996; Czypinski et al. 1997). The report provides an extensive table identifying the number and species of every fish collected. Detailed information regarding fish collected at each of the sites is found in the individual reports and will not be presented within the scope of this report. Several survey locations from the upper lakes that have been important in the history of the ruffe invasion are presented in this paper, as an example of how managers might use this information, especially when examining fish composition prior to or early in the invasion. Since bottom trawling is the primary method of collection, all of the data presented was selected from trawl efforts.

    The Bad River in western Lake Superior has been sampled annually since the program was initiated in 1992 (Figure 1). Ruffe were first collected there in 1993, marking the extent of its eastward expansion through that year (Figure 4). However, since only a small number of ruffe have been collected at the site each year, totaling only 13 ruffe since 1993, the Bad River continues to be sampled as part of the Program.

    Figure 4: Mean catch per unit effort (#fish/minute trawling) for the most commonly collected species in the Bad River, 1992-1995. Three letter abbreviations for common names are used to indicate species on the x-axis. The collection of ruffe is indicated by black bars.

    Surveillance has also been conducted in the Ontonagon River, Lake Superior, since 1992 (Figure 1). However, ruffe were not collected there until 1994 (Figure 5). This finding, together with the collection of ruffe in the Black River, Lake Superior, the same year, marked the first collections of ruffe in Michigan waters. The single ruffe collected in the Ontonagon River marked the farthest range expansion documented in a single year. In addition, the expansion occurred over a geographic area of Lake Superior's southern shoreline possessing some of the most unfavorable ruffe habitat. It had previously been hoped that the unsuitable habitat might provide a natural barrier slowing expansion. Although only one other ruffe was collected in 1995 surveys, the sighting causes additional concerns because of its close proximity to the entrance of the Keweenaw Waterway, the pathway to eastern Lake Superior.

    Figure 5: Mean catch per unit effort (#fish/minute trawling) for the most commonly collected species in the Ontonagon River, 1992-1995. Three letter abbreviations for common names are used to indicate species on the x-axis. The collection of ruffe is indicated by black bars.

    Surveillance efforts were initiated in Thunder Bay, Ontario, by the OMNR in 1991 to verify the presence of ruffe in the harbor following the capture and reporting of a single ruffe by an angler (Figure 1). Although ruffe were confirmed in trawl catches in 1991, they were not caught again in the Bay area until 1994, and then again in 1996 (Figure 6). Six ruffe were collected in 1996, amounting to a total of 15 ruffe collected from Thunder Bay. The need to continue surveillance seems apparent as numbers have remained very low over the six years since they were first found.

    Figure 6: Mean catch per unit effort (#fish/minute trawling) for the most commonly collected species in Thunder Bay, 1991-1996. Ruffe (black bars) were detected in surveillance efforts 1991, 1994, and 1996.

    Mail Survey
    To increase the comprehensiveness of the Program, both agencies recognized the role that public and private agencies, organizations, or individuals could play in surveillance efforts. In 1994, the OMNR initiated a program expanding its surveillance capabilities. A voluntary program was established targeting agencies, organizations, or individuals that routinely collect or handle Great Lakes fish or that expect to conduct fishery surveys in the Great Lakes. In 1995, the Service initiated a similar program, lead by the Ashland Fishery Resources Office.

    Each year potential participants are contacted twice through the mail. The first mailing is an informational packet that includes educational materials such as brochures, pamphlets, and identification cards regarding the ruffe. An explanation of the program, the role of participants in surveillance and the importance of documenting range expansions is included. The first mailing also encourages participants to report any potential ruffe sightings immediately to the Service or to OMNR. In the fall, following the field season, a follow-up questionnaire is mailed encouraging participants to provide information regarding their sampling efforts and fish collection. The information requested is simple, including sampling location, collection method/gear, a general approximation of the number of fish handled, and the presence or absence of ruffe.

    To provide some insight to the scope of this component, in 1995 the Ashland Fishery Resources Office sent out 439 packets to agencies and tribal commercial fishermen in the upper lakes. Response rates from both Canadian and U.S. participants have been low. However, even with low response rates, overall surveillance within the Great Lakes is increased substantially through this initiative. For example, Canadian responses in 1996 expanded reporting on the presence or absence of ruffe to Canadian waters of each of the Great Lakes (Figure 7). Many of the sites are included in other sampling efforts conducted by OMNR or the Canadian Department of Fisheries and Oceans, such as sea lamprey assessments or control treatment assessments, annual standard nearshore and offshore assessments, and creel surveys. Industries have been cooperative as well, providing information on species collected on intake screens, trash racks, etc.

    Figure 7: Locations included in surveillance efforts through responses to OMNR's Mail Survey questionnaire. This component has expanded their program to Canadian waters throughout the Great Lakes. (9. refers to the Invading Species Hotline established by the Ontario Federation of Anglers and Hunters.)

    Public Education
    The public has historically played an important role in nonindigenous species prevention activities, especially surveillance. An informed public is critical to the success of nonindigenous species activities in general. More specifically, surveillance programs may be enhanced significantly through the participation of public water users. In several cases, recreational anglers were the first to report the presence of ruffe in new locations, findings that were then verified by surveillance crews. A public that is aware of newly introduced species, pathways of introduction, and what to do if they think they have caught a potential candidate, will contribute to surveillance efforts as well as to control or prevention efforts minimizing the risk of further introductions or range expansions. Education initiatives are the primary tools managers possess to reach public water users.

    Working with Sea Grant agencies in the Great Lakes states, both the Service and OMNR have developed or assisted in the development of various educational materials targeting anglers, bait dealers and harvesters, marina owners, etc. Posters, brochures, and pamphlets have been distributed to anglers and angling groups, marinas, bait shops, and public access sites throughout the Great Lakes. Development of the Ruffe WATCH identification card was likely one of the most successful education initiatives. This wallet-sized card was designed so that anglers could easily store them in tackle boxes, with other fishing gear, or with their licenses. It was produced in a number of formats creating regional cards. Each card includes regional contact agencies, phone numbers for reporting potential sightings, and information regarding what to do with the candidate fish. The need for strong education programs and the support anglers may provide to surveillance has been proven, as so often anglers are the first to report sightings.

    Conclusion
    The Great Lakes Ruffe Surveillance Program provides resource managers with a model response to the introduction of nonindigenous species. It exemplifies an effective means to assist managers in 1) identifying newly established populations early, 2) detecting range expansion, 3) estimating potential impacts of introductions or range expansions by gathering baseline data, and 4) evaluating control or management strategies. As shown through this overview, the effectiveness and scope of surveillance programs can be enhanced significantly by conducting a multi-agency program consisting of components that broaden surveillance capabilities. This Program, conducted by the Service and the OMNR consists of three components: field surveys, a mail survey, and public education. The implementation of these components by both agencies has expanded surveillance efforts to U.S. and Canadian waters in all five of the Great Lakes.

    The field survey component assists managers to achieve all four of the above objectives. Field surveys conducted along the periphery of the ruffe's established range as well as in areas where ruffe are likely to become established, are apparently capable of detecting ruffe, even at relatively low population densities. This is likely due to the ruffe's high selectivity for favorable habitat at low densities. This ability is critical to early detection. However, managers must accept the limitations of sampling gear, realizing that absolute certainty can not be achieved (i.e. just because ruffe were not collected in a given area, does not mean that they are not there).

    The detection of new populations or range expansions inherently provides managers with a mechanism to evaluate control measures. For example, only two range expansions of ruffe within the Great Lakes are believed to have been associated with something other than natural movement. Intra- and inter-lake ballast transfers are likely responsible for introducing ruffe to Thunder Bay, Ontario, and Alpena, Michigan. However, only the Alpena, Michigan, sighting was documented following the implementation of the Great Lakes Maritime Industry Voluntary Ballast Water Management Plan for the Control of Ruffe in Lake Superior Ports in 1993. In 1995, when field survey crews first found ruffe in Alpena, Michigan, age determination indicated that ruffe had likely been introduced into the area prior to the implementation of the ballast water control strategy.

    The field survey component also provides an annual, consistent sampling program, from which fish species composition and abundance, habitat and water quality parameters can be estimated. This provides resource managers with information regarding the benthic fish community and habitat in areas where previously little to no information may have existed. Obtaining this type of information prior to the establishment of ruffe will assist managers with estimating potential impacts to the system following their invasion.

    The mail survey and public education components of the Program broaden surveillance capabilities. The amount of staff and time that can be dedicated to any fisheries program is limited. By educating other water users, surveillance is expanded beyond those limitations. Mail surveys target agencies, organizations, and individuals that routinely collect Great Lakes fish or that expect to conduct fishery surveys in the Great Lakes. This program includes government agencies, academic researchers, commercial fishermen, industries that draw water, etc. In addition, the public education component targets all potential water users. The value of public education has been proven, as anglers are often the first to report new sightings of nonindigenous species.

    This Program exemplifies a pro-active management approach to nonindigenous species. The information obtained through the Program has played and will continue to play a significant role in policy- or decision-making processes.

    Acknowledgements
    The authors would like to thank Gary Czypinski for providing data used in preparing graphs for this document and Amy Benson of the U.S. Geological Survey - Biological Resources Division for assistance in providing distribution and survey site maps. The authors are grateful to all staff and volunteers who have assisted with conducting field surveys, fish identification, and data entry. In addition, the authors would like to thank all those who have responded to the mail surveys and anglers who have reported the collection of potential nonindigenous species, including the ruffe.

    References
    Busiahn, T.R. 1997. Ruffe control: A case study of an aquatic nuisance species control program. In: Zebra Mussels and Aquatic Nuisance Species. F.M. D'Itri ed., Ann Arbor Press, Inc., Chelsea, MI, pp 69-86.

    Busiahn, T.R. and J.R. McClain. 1995. Status and control of ruffe (Gymnocephalus cernuus) in Lake Superior and potential for range expansion. In: The Lake Huron Ecosystem: Ecology, Fisheries and Management. M Munawar, T. Edsall, and J. Leach eds., SPB Academic Publishing, Amsterdam, The Netherlands, pp 461-470.

    Czypinski, G. D., G. Johnson, A.K. Hintz, and S.M. Keppner. 1997. Surveillance for Ruffe in the Great Lakes, 1996. U.S. Fish and Wildlife Service, Ashland Fishery Resources Office, Ashland, WI.

    Dermott, R. 1997. Changing amphipod community in the lower Great Lakes following the introduction of Ponto-Caspian species. Seventh International Zebra Mussel and Aquatic Nuisance Species Conference Program and Abstracts, January 28-31, 1997, New Orleans, LA. pp 40.

    Kindt, K., S.M. Keppner, and G. Johnson. 1996. Surveillance for Ruffe in the Great Lakes, 1995. U.S. Fish and Wildlife Service, Ashland Fishery Resources Office, Ashland, WI.

    Mills, E.L., J.H. Leach, J.T. Carlton, and C.L. Secor. 1993. Exotic species in the Great Lakes: A history of biotic crises and anthropogenic introductions. Journal of Great Lakes Research 19(1):1-54.

    Pratt, D.M., W.H. Blust, and J.H. Selgeby. 1992. Ruffe, Gymnocephalus cernuus: Newly introduced in North America. Canadian Journal of Fisheries and Aquatic Sciences 49:1616-1618.

    Simon, T.P. and J.T. Vondruska. 1991. Larval identification of the ruffe, Gymnocephalus cernuus (Linnaeus) (Percidae: Percini), in the St. Louis River Estuary, Lake Superior drainage basin, Minnesota. Canadian Journal of Zoology 69:436-442.

    Slade, J. and K. Kindt. 1992. Surveillance for Ruffe in the Upper Great Lakes, 1992. U.S. Fish and Wildlife Service, Ashland Fishery Resources Office, Ashland, WI.

    Slade, J.W., S.M. Pa, and W.R. MacCallum. 1994. Surveillance for Ruffe in the Great Lakes, 1993. U.S. Fish and Wildlife Service, Ashland Fishery Resources Office, Ashland, WI.

    Slade, J.W., S.M. Pa, and W.R. MacCallum. 1995. Surveillance for Ruffe in the Great Lakes, 1994. U.S. Fish and Wildlife Service, Ashland Fishery Resources Office, Ashland, WI.

    Zaranko, D. 1996. New exotic found in Lake Ontario. Great Lakes Commission Advisor. 9(4):7.

    Round Gobies: Cyberfish of the Third Millennium
    David J. Jude, Research Scientist
    Center for Great Lakes and Aquatic Sciences
    2200 Bonisteel Blvd.
    University of Michigan
    Ann Arbor, MI 48109-2099 USA

    Introduction
    The problem of transport of non-indigenous species from one place to another transcends boundaries, jurisdictions, and most of the efforts of humans to contain them. A recent study by Mills et al. (1995) lists 139 species that have made a substantial impact on Great Lakes ecosystems and they include organisms from bacteria, such as furunculosis, through purple loosestrife (Lythrum salicaria) and Eurasian watermilfoil (Myriophyllum spicatum), to fish. The transfer of these organisms was brought dramatically to the attention of businesses and the public with the discovery of zebra mussels (Dreissena polymorpha) in 1986 (Hebert et al. 1989) as water intakes became clogged, power plant pipes became plugged, and boats and motors became covered with this species. The homogenization of our aquatic communities, loss of biodiversity, and amalgamation of our gene pools because of the introduction of exotic species is a worldwide problem of which the round goby (Neogobious melanostomus) is just another symptom. The round goby along with its small cousin, the tubenose goby (Proterorhinus marmoratus), are our latest uninvited piscine immigrants joining Great Lakes fish communities. These fish are cyberfish, because they came from a distant universe and have the unusual ability to attain high abundances in optimal rocky substrate areas in the face of native fish communities and they also are able to disperse rapidly using Great Lakes freighters as transport vectors.

    Figure 1: Drawings of the round goby (Neogobius melanostomus) showing the fused pelvic fin and the dorsal black spot, and the tubenose goby (Proterorhinus marmoratus) showing the prominent nostrils.

    Life History
    The tubenose goby is a member of the family Gobiidae, whose members possess a unique fused pelvic fin (suctorial disk); this character can be used by the discerning individual to separate these fish from all other fishes in the Great Lakes basin (Fig. 1). It also has elongated nostrils, which are characteristic of this species. The tubenose goby is a small fish reaching lengths up to 110 mm (3 in), lives to about 5 years, and once the males spawn they die. They defend nests under rocks, shells, logs, and produce young that resemble adults (Miller 1984b). The diet of small gobies consists of aquatic insects and benthic zooplankton (Jude et al. 1995); some specimens we collected ate round goby fry in the St. Clair River (Jude, unpublished data). The tubenose goby has only been found in the St. Clair River, Lake St. Clair, and within the first 1 km of the Detroit River. It has not spread as far or as fast as the round goby. Population abundances would be considered low and they appear to have had negligible impact on fish communities to date. Ironically, because of habitat destruction, the tubenose goby is endangered in its native habitat of the Black and Caspian Seas in Russia (Lelek 1989).

    The round goby also has the characteristic fused pelvic fins, elongated dorsal and anal fins like the tubenose goby, and a unique black spot on the dorsal fin. Interestingly, a large percentage of the round gobies in Lake Erie do not have this black spot, which as far as we can tell, has not been observed in Europe. The round goby attains a much larger size, up to 250 mm or 10 in, and lives for up to 5 years (Berg 1949). After attaining a large size, males spawn once, then normally die. The keys to why certain males will defend a nest and others do not is unknown. It may be related to finding a suitable site, size, or aggressiveness. Round gobies are known to produce several vocalizations, including one that attracts females to nest sites and one that intimidates (a growling sound) males (Protasov et al. 1965). Females, however, do not die after spawning, and can spawn up to six times during the spring and summer, every 20 days conveying a great competitive advantage over native species that usually spawn once during spring. Males turn coal black and defend nests made under rocks, in beer cans, and in hollowed out logs. Large eggs (fecundity is around 5,000 eggs/female - Miller 1984a) are deposited on the undersides of these structures; several females can deposit distinctive, cone-shaped eggs at one site (Miller 1984b). Round goby larvae also resemble adults at hatching and appear to be benthic, since they have no swim bladder. The round goby is an important component of the commercial catch and is eaten regularly in Europe, but has declined precipitously in recent years because of degradation of habitat (Moyle and Cech 1988).

    The round goby is robust, aggressive, and can withstand very low levels of dissolved oxygen (Jude unpublished data) and apparently defends vigorously the type of rocky and cobble habitat it prefers to the exclusion of some native species. Small round gobies were most common in the nearshore rocky riprap area, but eggs were trawled from deeper water. Round gobies were also found in beach areas with predominantly sandy substrate, but numbers were reduced in these areas. They were most prominent in depths of 3-5 m, but were found all the way out to 10 m in the channel of the St. Clair River (Jude et al. 1995). There appeared to be a direct relationship between length of round goby and depth where fish were found. Round gobies are adapted to be primarily mollusk eaters, which their well developed, molariform teeth attest. The studies we have conducted so far indicate that at small sizes they eat benthos and benthic zooplankton while an increasing proportion of their diet as they grow older is zebra mussels, fingernail clams, and snails. Most zebra mussels eaten were 2-12 mm in length (Jude et al. 1995). Ghedotti et al. (1995) has recorded the consumption of over 100 zebra mussels per day in laboratory studies. Stomachs from large individuals we examined generally contained 10-12 large zebra mussels from 6-12 mm.

    Figure 2: Current distribution of the round goby in the Great Lakes (black dots), including one record in central Michigan in the Shiawassee River. The tubenose goby is found in the St. Clair and Detroit Rivers and Lake St. Clair

    Round gobies appear to be adapted to warm temperatures and growth of our fish is much slower and maximum lengths are smaller than specimens from the Black and Caspian Seas. In addition, fish from Lake Erie, which is presumably much more productive and warmer than Lake Michigan or the St. Clair River grew the fastest among Great Lakes round gobies we examined (Jude et al. unpublished data). Dougherty et al. (1995) has found that Great Lakes specimens we supplied from the St. Clair River, and that he examined for mitochondrial DNA were not derived from native populations he sampled from the Black Sea near Varna, Bulgaria. A hybrid species between N. melanostomus and fluviatilis has been documented in Europe (Pinchuk 1970).

    Distribution
    First discovered in 1990 in the St. Clair River (Jude et al. 1992), the round goby has now spread throughout the St. Clair River system into Lake St. Clair, the head of the Detroit River, Lake Erie, southern Lake Michigan, southern Lake Huron, Lake Superior, and there are recent reports of their capture in Lake Ontario, completing their dispersal throughout the five Great Lakes in five years, dispersal almost as fast as the zebra mussel (Fig. 2). It took the alewife (Aloso pseudoharengus) and sea lamprey (Petromyzon marinus) almost 25 years to make it into Lake Superior after once reaching Lake Erie (Jude and Leach 1994). The early non-indigenous species probably moved from lake to lake by swimming, while round gobies and ruffe (Gymnocephalus cernuus) (Pratt et al. 1992) were ballast water transfers and continue to be transferred throughout the Great Lakes by the same mechanism. In addition, there is one known occurrence of round gobies inland in the Shiawassee River, near Fenton, Michigan (Raesly, R., personal communication, Frostberg State University, Frostburg, Maryland).

    Rela and Potential Impacts
    Predation by round gobies on other fish has only been recorded once, when a 41-mm adult trout-perch (Percopsis omiscomaycus) was found inside a 139-mm round goby. Round gobies have been caught by fishermen using minnows; they have killed cohabitating darters (Etheostoma) in confined aquaria (probably a territorial response); and are known to eat fish in their native habitat (Miller 1986). This opens up another avenue of potential impact of round gobies on native species, since round gobies can prey on fairly large, cohabitating benthic fishes.

    Competition for limited resources at early life stages may be another interface of interaction between round gobies and other native fish. Round gobies attain high abundances and eat benthic zooplankton and benthos which is important food for the young of cohabitating mottled sculpins (Cottus bairdi), darters, logperch (Percina caprodes), and some minnows (Notropis spp.). Diet overlap with some of these species is substantial (French, J., D. Jude, and G. Crawford, unpublished diet data).

    Our research in the St. Clair River has indicated that round gobies have seriously impacted, a native, bottom-dwelling species, the mottled sculpin, causing a dramatic decline in their populations from when they were the second- or third-most abundant species in previous studies done in the river. The remnant of the surviving mottled sculpin population was forced into deeper water, where the round goby appears to be less abundant. We hypothesize that the round goby, because of differences in the neural canals between these two species, does not compete as well with the mottled sculpin in deep, high velocity waters, as it obviously does in shallow water (Jude et al. 1995). In shallow water, the mechanism of this competitive advantage of round gobies over native species appears to be the aggressive nature of the round goby which drives mottled sculpin from prime feeding, shelter, and especially spawning areas. We think that mottled sculpin spawning has been disrupted causing almost total year class failure of the mottled sculpin in areas where round gobies are abundant. This hypothesis is supported by evidence that shows that round gobies would drive mottled sculpin from one shared shelter in our laboratory aquarium studies and recent studies by Dubs and Corkum (1996). In addition, field studies conducted by John Janssen, Loyola University were initiated in the Calumet River Harbor area in southern Lake Michigan before round gobies became abundant in that area in 1995. It is clear from his SCUBA observation data (Fig. 3) that mottled sculpins suffered year class failure starting in 1995 (when round gobies started to become common), since no YOY mottled sculpin were observed there despite their presence in 1994, along with observations of gravid females. In 1996, not only YOY were missing from the age-groups, similar to what happened in 1995, but yearlings were also absent. In 1996, again no YOY mottled sculpin were found, however, there were several round goby defended nests present in the area. Very few mottled sculpins were observed, except for old adults apparently coming in from other areas. This pattern has therefore been repeated in two areas of high abundance of round gobies.

    There are three possibly damaging scenarios we are concerned about that may come to pass in the future. The first of these is food chain bioaccumulation of PCBs. Since zebra mussels are known to accumulate large concentrations of PCBs (deKock et al. 1993), and round gobies eat large numbers of zebra mussels, and now we have sport fish eating round gobies (Jude et al. 1995), the potential for contamination of important sport fish is of obvious concern.

    Secondly, another species of goby, the black goby (Gobius niger), has been documented in a Netherlands lake to have caused the demise of a species of sculpin, which is a congener of our Great Lakes deepwater sculpin (Vass et al. 1975). The deepwater sculpin is very abundant in the deep abyss of the upper three Great Lakes. The potential exists for the round goby to have a negative impact on these populations. The literature states that round gobies will overwinter in water up to 60 m deep in the Black Sea (Miller 1986). Deepwater sculpin usually reside in water 50 m or deeper. To date we have evidence that in the Great Lakes, round gobies have been found in 18 m (60 feet) of water in Lake Erie, 15 m (50 feet) of water in Lake Michigan and in 10 m of water in the St. Clair River. Therefore, the potential definitely exists for this species to occupy deeper water and affect deepwater sculpin, either by competing with them for food (Mysis and Diporeia) or eating eggs from their nests. Deepwater sculpin make shallow depressions on the bottom during winter, eggs are laid, and then guarded by the male. However, the colder water there (constant 4( C) may inhibit round gobies from detrimentally affecting deepwater sculpin.

    Third, round gobies are currently confined to the Great Lakes and connecting channels except for one occurrence in the Shiawassee River in mid-Michigan. Round gobies have penetrated about 19 km upriver in the Calumet River system in southern Lake Michigan, which is about 480 km away from the Mississippi River (Steingraeber et al. 1996). Once they reach the Mississippi River or some inland lakes, round gobies will be exposed to another new ichthyofauna and we know nothing about those interactions. Their current ability to depress mottled sculpin populations, a nest guarder, makes us concerned about interactions with very important native nest guarders, such as the bluegill and large and smallmouth basses.

    Ballast Water Transport
    There have been a number of ways in which exotic species have entered the Great Lakes, with ballast water discharge by oceangoing vessels recently attaining prominence as the leading causes of these unwanted introductions (Mills et al. 1995, Carlton 1985, Moyle 1991). The presence of round gobies in the St. Clair River is a symptom of our lack of dealing with the global transport of non-indigenous species. The recent examples of zebra mussels, ruffe, and the gobies are testimony that we still have not curtailed this threat to our aquatic ecosystems. Further evidence of our lack of control of this problem is that U.S. Coast Guard mandatory and Canadian voluntary restrictions on foreign freighters to exchange ballast water with seawater in order to kill freshwater organisms being transported were just beginning at the time of most of these latest introductions. Obviously more must be done, and there are in fact recent efforts to design a pilot filtration apparatus aboard the freighter Algonorth to evaluate the efficacy of this approach of treating ballast water.

    Round gobies may be pre-adapted for transport in the ballast water of transoceanic ships and also freighters within the Great Lakes, which is probably how they got here in the first place and probably why round gobies have been able to spread so far so fast. The family Gobiidae is the family with the second largest number of species of fish, some 2,000 species, and this includes many well-adapted species, such as cavefish, which live in complete darkness, not unlike the ballast water environment of a ship. Our new found goby guests should not come as a surprise, nor should the introduction of additional species in this family. There are many incidences of freighter transport of other species of gobies into other water bodies across the world, including new introductions into Australia, San Francisco Bay, the Arabian Sea, and recently the round goby has extended its range into the Aral Sea and into the Baltic Sea in the Bay of Gadansk, Poland (Moyle and Cech 1988, Hoese 1973, Al-Hassan and Miller 1986, Miller 1984b, Skora and Stolarski 1996). The patterns we have seen in the Great Lakes, rapid expansion in areas of optimal rocky habitat, have been repeated in these environments as well. So our experiences are not unique.

    Figure 3: Length-frequency distribution of mottled sculpins during 1994-1996. Fish were observed during SCUBA surveys at one site in Calumet Harbor, southern Lake Michigan. Note the progressive loss of young-of-the-year and yearling fish as round gobies became abundant in 1995

    There are many reasons why round gobies continue to be carried around the Great Lakes in ballast water of ships. First, ships routinely take on large quantities of water, thousands of tons, from a place like the St. Clair River where round gobies are very abundant. This water is transported elsewhere and discharged. It is no wonder that the major ports in the Great Lakes, Duluth Superior, Calumet River, and the Grand River in Lake Erie were sites where they were initially dumped and have subsequently done very well. These river mouths and harbor areas have extensive areas of riprap, sheet piling, and other structure that provides ideal substrate for hiding, feeding on zebra mussels, and for reproductive activities. Round gobies are very cryptic species; they occupy holes, crevices, and cracks and may have occupied these places on ships. They may also have laid eggs on the undersides of ships or in suitable holes thereby allowing the eggs to be transported long distances before they hatched.

    Round gobies have a lateral line system (neuromasts in the head region) which is very well adapted for feeding in the dark and is almost unique among Great Lakes species. Studies performed by John Janssen on round gobies have shown that round gobies can feed in complete darkness and can detect prey at a shorter distance than can mottled sculpin, giving them a competitive advantage in this area as well (Jude et al. 1995). This creature feature, along with the round gobies' tolerance for degraded water quality, has probably pre-adapted round gobies for transport in ships and their ballast water.

    Predators
    Predators have shown an interest in round gobies in the St. Clair River, probably because they are so abundant in that system. Predators we found to have preyed on round gobies included rock bass (Ambloplites rupestris), smallmouth bass (Micropterus dolomieui), stonecat (Noturus flavus), tubenose goby, brown trout (Salmo trutta), walleye (Stizostedion vitreum), yellow perch (Perca flavescens), rainbow smelt (Osmerus mordax), and even mottled sculpin. Cannibalism was also common for this species.

    Reasons for Dominance of Aquatic Habitats and certain fishes
    Round gobies have dominated optimal rocky habitat areas in the Great Lakes (Jude and DeBoe 1996) because they grow to a larger size than cohabitating native benthic fishes, like the darters, logperch, and mottled sculpin. In addition, they are aggressive, spawn often during the warm water months producing many cohorts (some of which are guaranteed to survive), guard the nests insuring high survival of young, and they can feed in complete darkness. Another reason that round gobies have been so successful can be attributed to zebra mussels. It is known that zebra mussels transfer energy from the pelagic zone (water column) to the benthic zone (Klerks et al. 1996) through their filtering capabilities, fueling a large growth of zebra mussels at the expense of zooplankton which would have fed on the algae and detritus removed by the mussels (Dermott et al. 1992, Griffiths 1992). Zebra mussels are not often eaten by native fish species (French and Love 1995), so they represent a dead end ecologically. However, round gobies greater than 100 mm feed almost exclusively on zebra mussels with little competition, allowing them to build up large populations, which can cause problems because of sheer numbers alone (Jude et al. 1995). For example, we often got over 200 round gobies in 10-min hauls with our 5-m headrope trawl, and up to 500 per haul were collected in Lake Erie. In addition, since the invasion of the quagga mussel (Dreissena bugensis), which has a cooler and deeper water tolerance (Rosenberg and Ludyanskiy 1994; Spidle et al. 1994), another source of mussels may be available for round gobies in the deeper, soft sediment areas of the Great Lakes where zebra mussels were prevented from colonizing because of the lack of hard substrate. This is especially true for Lake Erie and Ontario, and may also allow quagga mussels to flourish in Lake Superior, where zebra mussels have not done very well.

    Conclusions
    Appearance and dispersal of round gobies, which have become a permanent member of Great Lakes fish communities, are causes of loss of biodiversity and a symptom of our damaged ecosystems in the Great Lakes, since entry of exotic species are favored when fish community integrity is severely destabilized. Exotic species, including the introduced salmon (Oncorhynchus spp.) and rainbow (O. mykiss) and brown trout, alewives, rainbow smelt, and the infamous sea lamprey have irrevocably changed ecosystem form and function in the Great Lakes. The round goby is just another addition to our growing list of exotic species of not only fish, but of other groups, including invertebrates, well represented by the zebra mussel, and plants, including the notorious purple loosestrife, which is destroying our wetlands. These organisms are wreaking more havoc with ecosystem health than toxic substances or nutrients, since we at least have done, and can continue to do, something about controlling those problems. Once in, most successful exotic species establish permanent populations which are irrevocably detrimental to fish community stability and cause serious loss of irreplaceable genetic material and biodiversity. Their effects are difficult to deal with, complex, and like a broken heart on Valentines Day, it is something that we just can't fix!

    Literature Cited
    Al-Hassan, L. A. J., and P. J. Miller. 1987. Rhinogobius brunneus (Gobiidae) in the Arabian Gulf. Japanese Journal of Ichthyology 33(4):405-408.

    Berg, L. S. 1949. Freshwater fishes of the USSR and adjacent countries. Vol. III, I-2d-vo. AN SSR 510 pp (translated by Israel Program for Scientific Translation, Jerusalem 1965).

    Carlton, J. T. 1985. Transoceanic and interoceanic dispersal of coastal marine organisms; the biology of ballast water. Oceanogr. Mar. Biol. Ann. Rev. 23:313-371.

    deKock, D., W. Chr., and C. T. Bowmer. 1993. Bioaccumulation, biological effects and food chain transfer of contaminants in the zebra mussel (Dreissena polymorpha), Chapter 30, pp. 503-533. In: T. F. Nalepa and D. W. Schloesser (eds), Zebra Mussels: Biology, Impact and Control. Lewis Publisher/CRC Press, Inc. Boca Raton, Florida.

    Dougherty, J.D., W. S. Moore, and J. L. Ram. 1996. Mitochondrial DNA analysis of round goby (Neogobius melanostomus) and tubenose goby (Proterorhinus marmoratus) in the Great Lakes. Can. J. Fish. Aquat. Sci. 53:474-480.

    Dermott, R., J. Mitchell, I. Murray, and E. Fear. 1992. Biomass and production of zebra mussels (Dreissena polymorpha) in shallow waters of northeastern Lake Erie. pp. 399-414. In: T. F. Nalepa and D. W. Schloesser (eds), Zebra Mussels: Biology, Impact and Control. Lewis Publisher/CRC Press, Inc. Boca Raton, Florida.

    Dubs, D. O.L, and L. D. Corkum. 1996. Behavioral interactions between round gobies (Neogobius melanostomus) and mottled sculpins (Cottus bairdi). J. Great Lakes Res. 22:838-844.

    French III, J. R. P., and J. G. Love. 1995. Size limitation on zebra mussels consumed by freshwater drum may preclude the effectiveness of drum as a biological controller. Freshwater Ecology 10(4):379-383.

    Ghedotti, M., J. Smihula, and G. R. Smith. 1995. Zebra mussel predation by round gobies in the laboratory. J. Great Lakes Res. 21:665-669.

    Griffiths, R. 1992. Effects of zebra mussels (Dreissena polymorpha) on the benthic fauna of Lake St. Clair. pp. 415-438. In: T. F. Nalepa and D. W. Schloesser (eds), Zebra Mussels: Biology, Impact and Control. Lewis Publisher/CRC Press, Inc. Boca Raton, Florida.

    Hebert, P.D. N., B. Muncaster, and G. L. Mackie. 1989. Ecological and genetic studies on Dreissena polymorpha (Pallas): a new mollusc in the Great Lakes. Can. J. Fish. Aquat. Sci. 46:1587-1591.

    Hoese, D. F. 1973. The introduction of the gobiid fishes Acanthogobius flavimanus and Tridentiger trigonocephalus into Australia. Kookewong 2(3):3-5.

    Jude, D. J., and J. Leach. 1993. Chapter 22: Great Lakes Fisheries. In: Inland Fisheries Management in North American, (Eds.) C. Kohler and W. Hubert, American Fisheries Society, Bethesda, Maryland, pp. 517-551.

    Jude, D. J., and S. DeBoe. 1996. Possible impact of gobies and other introduced species on habitat restoration efforts. Can. J. Fish. Aquat. Sci. 53(Suppl. 1):136-141.

    Jude, D. J., J. Janssen, and G. Crawford. 1995. Ecology, distribution, and impact of the newly introduced round and tubenose gobies on the biota of the St. Clair and Detroit Rivers. pp. 447-460. In: M. Munawar, T. Edsall, and J. Leach (eds.). The Lake Huron Ecosystem: Ecology, Fisheries, and Management. Ecovision World Monograph Series, S. P. B. Academic Publishing, The Netherlands.

    Jude, D. J., R. H. Reider, and G. R. Smith. 1992. Establishment of Gobiidae in the Great Lakes basin. Can. J. Fish. Aquat. Sci. 49:416-421.

    Klerks, P. L., P. C. Fraleigh, and J. E. Lawniczak. 1996. Effects of zebra mussels (Dreissena polymorpha) on seston levels and sediment deposition in western Lake Erie. Can. J. Fish. Aquat. Sci. 53:2284-2291.

    Lelek, A. 1989. Threatened Fishes of Europe. Vol. 9 in: European Committee for the Conservation of Nature and Natural Resources - Council of Europe (ed). The Freshwater Fishes of Europe. AULA-Verlag Wiesbaden.

    Mills, E., J. H. Leach, J. T. Carlton, and C. L. Secor. 1993. Exotic species in the Great Lakes: a history of biotic crises and anthropogenic introductions. J. Great Lakes Res. 19:1-54.

    Miller, P. J. 1984a. Introduction: strategies and tactics in fish reproduction. pp. 45-53 In: Wootton, R. J. (ed.). Fish Reproduction. Academic Press, Ltd., London.

    Miller, P. J. 1984b. Tokology of gobies, pp. 119-153. In: Wootton, R. J. (ed.). Fish Reproduction. Academic Press, Ltd., London.

    Miller, P. J. 1986. Gobiidae, P. 1019-1095. In Whitehead, P.J. P., M.-L. Bauchot, J.-C. Hureau, J. Nielsen, E. Tortonese, Eds., Fishes of the North-east Atlantic and the Mediterranean, Vol. III., UNESCO, Paris. Moyle, P. B., and J. J. Cech, Jr. 1988. Fishes: An introduction to ichthyology. Prentice Hall, Englewood Cliffs, New Jersey. 559 pp.

    Moyle, P. B. 1991. Ballast water introductions. Fisheries 16(1):4-6. Pinchuk, V. I. 1970. An unusual goby specimen from the bay of Taman. J. Ichthyology 1970 (10) vol. 5:689-691.

    Pratt, D. M., and W. H. Blust. 1992. Ruffe, Gymnocephalus cernuus: newly introduced in North America. Can. J. Fish. Aquat. Sci. 49:1616-1618.

    Protasov, V. I, V. I. Tzvetkov, and V. K. Rashchperin. 1965. Acoustic signalization of round goby (Neogobius melanostomus) from the Azov Sea. Journal of General Biology (Russian) 26,2:151-160.

    Rosenberg, G., and M. L. Ludyanskiy. 1994. A nomenclatural review of Dreissena (Bivalvia: Dreissenidae) with identification of the quagga mussel as Dreissena bugensis. Can. J. Fish. Aquat. Sci. 51:1474-1484.

    Spidle, A. P., J. E. Marsden, and B. May. 1994. Identification of the Great Lakes quagga mussel as Dreissena bugensis from the Dnieper River, Ukraine, on the basis of allozyme variation. Can. J. Fish. Aquat. Sci. 51:1485-1489.

    Skora, K.E., and J. Stolarski. 1996. Neogobius melanostomus (Pallas 1811), A new immigrant species in the Baltic Sea. pp. 101-108. In: Estuarine Ecosystems and Species, Ewa. Styczynska-Jurewicz (ED). Proceedings of the 2nd international Estuary Symposium held in Gdansk, 18-22 October 1993. Crangon Issues of the Marine Biology Centre in Gdynia No. 1, Gdynia, 1996.

    Steingraeber, M., A. Runstrom, and P. Theil. 1996. Round goby (Neogobius melanostomus) distribution in the Illinois Waterway system of metropolitan Chicago. U.S. Fish and Wildlife Service Publication, Fishery Resources Office, Onalaska, Wisconsin.

    Vass, D. F., A. G. Vlasblom, and P. De Koeijer. 1975. Studies of the black goby (Gobius niger, Gobiidae, Pisces) in the Veerse Meer, SW Netherlands. Netherlands Journal of Sea Research 9:56-68.

    Economic Impact of Zebra Mussels Results of the 1995 National Zebra Mussel Information Clearinghouse Study
    Charles R. O Neill, Jr.
    Coastal Resources Specialist
    New York Sea Grant
    Project Director, National Zebra Mussel Information Clearinghouse

    Introduction
    Raw water-dependent infrastructure, including electric power generation stations, drinking water treatment plants, industrial facilities, navigation lock and dam structures, and other facilities throughout much of the eastern half of North America, has been significantly impacted by zebra mussels. While the mussel's physical impacts upon infrastructure are well known, their economic impacts are less understood. This study, undertaken by New York Sea Grant and the National Zebra Mussel Information Clearinghouse, addressed the economic impact of the zebra mussel throughout the mussel's North American range, as well as regions proximate to that range but not yet known to be infested.

    A detailed survey was mailed to a random sample totaling of 766 infrastructure owners/operators throughout a 35 state / 3 province range. From this mailing, 436 usable responses were received, a 56.92% return rate. Three hundred thirty nine facilities reported expending funds related to zebra mussel impacts (see Map). Information solicited included: facility type and location; source water body; degree of facility water use; types of zebra mussel- related impacts; zebra mussel monitoring and control activities; and, 13 categories of zebra mussel-related annual costs from 1989 through 1995.

    Figure: First number represents facilities reporting zebra mussel-related economic impacts, second number represents number of zebra mussel infested facilities in that state

    Three hundred thirty nine facilities reported total zebra mussel-related expenses of $69,070,780, with a minimum reported expenditure of $400, a maximum expenditure of $5,953,000, and a mean expenditure of $205,570 per facility (see Table and Figure 1). The big spender was nuclear power plants with a mean expenditure of $786,670 per facility, accounting for 26.2% of the total reported zebra mussel impact. Other major water user categories included: drinking water treatment facilities, with a mean expenditure of $214,360 (31.03% of total reported impact); fossil fuel electric generating facilities, with a mean expenditure of $145,620 per facility (16.02% of the total reported impact); and, industrial facilities, with a mean expenditure of $167,030 per facility (8.46% of the total reported impact). Total annual expenditures increased from 1989 ($234,140) to 1995 ($17,751,000) as the mussel s North American range and the number of facilities affected increased (Figure 2).

    Figure 1: Summary of total zebra mussel economic impacts by water use category. Note lagarithmic scale.


    Figure 2: Total annual zebra mussel-related expenditures for all water user categories.

    Results
    No zebra mussel-related expenses were reported for non-navigation water transport canals, aquarium theme parks, or sewage treatment facilities. Two noninfested golf courses reported total monitoring expenses of $750 from 1993 through 1995. Two infested marinas, one on Lake Champlain and one on the Niagara River, reported total zebra mussel-related expenses of $7,500. One public recreation area on an infested public impoundment reported $20,000 of zebra mussel-related expenditures, mostly for monitoring and planning. Two institutions, a university campus on Lake Michigan and a residential care facility on Lake Champlain, reported total zebra mussel-related expenses of $27,000, $2,000 for monitoring and $25,000 for prevention. Nine noninfested public impoundments in three states reported total zebra mussel- related monitoring expenses of $27,100. Three fish hatcheries in three states reported zebra mussel-related expenses of $88,000 from 1993 through 1995, of which $73,000 was spent at two infested facilities. The uninfested facility uses ponds as its water source, one infested facility uses the Mississippi River, the other infested facility uses Lake Champlain. $67,000 was spent on preventive measures, $15,000 on research, $5,000 on planning and engineering, and $1,000 on monitoring.

    Responses were received from 58 public navigation lock facilities (28 of which are infested) in 10 states. The facilities are on the Allegheny, Arkansas, Cumberland, Mississippi, Monongahela, Ohio, and Verdigris Rivers and their tributaries. Fifty six of the facilities reported total cumulative zebra mussel-related expenses of $484,800 from 1991 through 1995, $441,700 (91.11%) at infested facilities, $43,100 at uninfested facilities. Navigation locks represent 13.9% of facilities responding to the survey, 15.7% of infested facilities, and 16.5% of facilities reporting zebra mussel-related expenses, yet locks accounted for only 0.7% of the total reported seven year zebra mussel economic impact. For this reason, any aggregate figure for all water user categories, including navigation locks, which shows an average per facility zebra mussel- related expense will be skewed toward the low end of the scale.

    Fifty six locks spent an average of $3,416 each on monitoring, ranging from $400 to $31,000, accounting for $191,300 (39.5%) of reported navigation lock expenses. The second largest expense category was facility retrofit, with two locks spending $100,000. One facility spent $75,000 on mechanical controls; two locks spent $36,000 on staff training; and three locks spent $34,000 on planning and engineering expenses. Several facilities were expected to close for several days in late 1995 for mechanical removal of mussels $32,000 revenue was expected to be lost during these closures. Other expenses included nonchemical controls, research, and prevention.

    One response was received from a shipping related entity with numerous vessels on the Arkansas, Hudson, Illinois, Mississippi, Ohio, Tennessee, and Missouri Rivers and the Great Lakes. Vessels were reported to have been infested since 1990; $563,000 worth of zebra mussel- related expenses were incurred in all categories. The single largest category of expense, $220,000 (39.08%), was for vessel retrofit, followed by $99,000 (17.58%) for nonchemical treatments and $92,000 (16.34%) for lost production/use. Other expense categories included: mechanical treatment, planning and engineering, training, research, prevention, monitoring, chemical controls, and filtration. It was reported that all new vessels are being constructed with zebra mussel-proofing measures built in.

    A National Scenic Riverway reported that it was not yet infested but was proactive in its response to zebra mussels, implementing a monitoring and user education programs, signage, boat inspections, and boat cleaning stations at access points. The facility spent $723,000 on monitoring, planning, prevention, research, and training activities from 1992 through 1995.

    Five government agencies (two federal, three state) verbally reported incurring $4,574,000 of zebra mussel-related expenses from 1991 through 1995, 6.62% of the total reported economic impact. In all cases, the money was spent on mussel control research.

    Fifty six industries in 13 states and one province responded, 22 (39.29%) of which reported that they are infested. No industrial expenses were reported for 1989. Thirty five (62.5%) of the industries reported a cumulative total zebra mussel-related expense of $5,846,000 from 1990 through 1995; $3,747,500 (64.1%) at infested facilities, $2,098,500 at uninfested facilities (Figure 3). Industrial expenses represented 8.46% of total reported zebra mussel economic impacts.

    Figure 3: Industrial facility expenses at infested and uninfested facilities (in $1000s) by year

    The largest number of responses from a single geographic region came from 18 facilities on the Mississippi River with a total expense of $753,000 (12.88% of industrial expenditures); seven facilities drawing their water from the Great Lakes reported $2,815,000 of expenditures (48.15%). This disparity in impact can be attributed to the Great Lakes facilities becoming infested much earlier than those on the Mississippi. Expenses generally increased from 1991, with the exception of 1992 which, at $1,423,500, was the second highest annual expenditure. 1992 was also the only year in which expenditures at uninfested facilities ($924,000) exceeded those at infested facilities ($499,500).

    Eight industries (five of which are infested) spent a total of $2,657,000 (a mean expenditure of $332,125 per facility) on preventive measures, accounting for 45.45% of the total reported industrial expenses, the largest single expense category (Figure 4). These expenditures, however, were very unevenly distributed, ranging from $4,000 to $1,500,000 at facilities with water use capacities ranging from 4 million gallons per day (mgd) to 300 mgd (mean capacity = 60.3 mgd). Excluding two facilities that plot as outliers, the mean industrial prevention expense at facilities with water use capacities between 4,000 and 50,000 mgd was $92,833.

    Figure 4: Industrial expenses by category (in $1000s).

    The second largest expense category was chemical control measures. Seventeen industries with water use capacities ranging from 4 to 300 mgd (mean capacity = 51.5 mgd) spent a total of $1,358,000 (mean expenditure of $79,882 per facility), 23.2% of the total industrial expenditure. Twenty one industries with water use capacities ranging from 3 to 94 mgd spent a total of $781,000 (mean expenditure of $37,190 per facility) on planning related expenditures. Twenty eight facilities spent a total of $401,500 (mean = $14,393) on monitoring activities. Five facilities spent a total of $241,000 (mean = $48,200) on retrofit of existing plant. Eight facilities spent an average of $20,250 for a total of $162,000 on mechanical control alternatives. Two facilities spent a total of $128,000 on other (thermal and coatings) control technologies. Sixteen facilities spent a total of $102,500 (mean = $6,406) on staff training, and two facilities spent a total of $15,000 on research activities. No funds were spent on lost production, filtration, or nonchemical, nonmechanical control techniques.

    Responses were received from 160 drinking water treatment plants (WTPs) in 30 states and two provinces. One hundred facilities, with a combined water usage of 5,677,440 mgd (62.5% of all WTPs responding to the survey) reported that they had spent a combined total of $21,445,610 on zebra mussel related activities, 31.05% of the total reported zebra mussel economic impact. The state with the greatest number of water treatment plants responding was Alabama with 27 responses, one of which was infested; New York had the second largest response, 21, of which 15 are infested. The states/provinces which spent the greatest amount of money on zebra mussel related activities at WTPs were: New York ($7.63 million, 35.6% of all WTPs), Maryland ($6.11 million, 28.5%), and Wisconsin ($2.43 million, 11.3%)(Figure 5). Water treatment plant expenses began in 1989 in the Great Lakes region, with seven facilities reporting a total expenditure of $106,140 compared to three non-Great Lakes facilities spending a total of only $16,000. Both the number of impacted plants and total expenditures increased on a yearly basis with the Great Lakes outpacing other regions of the country until 1994, when 49 non-Great Lakes facilities spent a combined total of $3,483,300 compared to 41 Great Lakes facilities spending $2,331,800. Over the entire period, Great Lakes facilities spent a total of $11,717,900 (54.64% of WTP expenditures), while non-Great Lakes facilities spent $9,727,700.

    Figure 5: Water treatment plant expenditures (by state/province, in $1000s), 1989-1995 inclusive. (Note: logarithmic scale)

    From 1989 through 1991 the majority of WTP expenditures were spent responsively by infested facilities. This changed in 1992 when proactive expenditures at noninfested facilities began to exceed those at infested plants. This trend was very evident from 1993 through 1995 with noninfested facilities out spending infested ones by 47% in 1993, 487% in 1994, and 244% in 1995. Thirty eight infested facilities (total water use of 932,900 mgd) reported that they spent a total of $6,526,910 (30.4% of WTP expenses), while noninfested facilities spent the remaining $14,918,700 (Figure 6).

    Figure 6: Water Treatment plant expenses at infested and uninfested facilities (in $1000s) by year

    Thirty five water treatment plants (combined total water use capacity of 908 mgd) spent a total of $6,385,570 on physical plant retrofit (29.78% of all WTP expenses), with a mean per facility expenditure of $182,440 (Figure 7). In all cases, this was to provide capability for preoxidation of intake water at the source water end of the facilities intakes. The second largest category was planning and engineering expenses, with 55 water treatment plants (combined total water use capacity of 4.091 mgd) spending a total of $6,229,450 (29.05% of all WTP expenses), with a mean per facility expenditure of $113,265. Thirty two facilities (combined total water use capacity of 36.17 mgd) spent $6,221,460 (29% of all WTP expenses) on chemical control activities, with a mean per facility expenditure of $194,420. Of the 60 facilities that reported using chemical treatment methods, all but two are using oxidizing chemicals, with various forms of chlorine making up 70% of the total, chlorine dioxide 15%, potassium permanganate 12%, ozone 2%, and bromine 1%. Two facilities reported using polyquaternary ammonium compounds. Twenty one facilities spent a total of $926,200 on prevention-oriented activities, and 80 facilities spent $914,780 on monitoring. Other important categories included: mechanical removal (hard-helmet divers and pigging, 15 facilities, $236,590), staff training (38 facilities, $161,760), nonchemical controls (filtration, 2 facilities, $142,500) other activities (use of polymers, plasma sparking, and alternating intakes, 6 facilities, $108,670), research (8 facilities, $98,630), and, lost revenue or production ($20,000 at 1 facility).

    Figure 7: Water Treatment plant expenses by category (in$1000s)

    One hundred thirty three responses were received from the electric generation industry, by far the largest single surface water user group to be economically impacted by zebra mussels, with 124 respondents reporting an industry-wide expenditure of $35,274,020 from 1989 through 1995, 51.07% of all reported zebra mussel-related expenses (Figure 8). The reason for this lies in the operational differences between this industry and other water users. A drinking water facility, for example, might treat two or three hundred million gallons per day, whereas a power plant, with its tremendous water needs for heat exchanger cooling, bearing lubrication, transformer cooling, and fire control, to use two or more billion gallons per day. Pumping so much water through power plant water systems brings those systems into direct contact with astronomical numbers of mussel larvae, juveniles and adults, and results in a greater potential biofouling impact. In response to such high levels of potential impact, the electric industry has expended higher levels of funds for research, monitoring, remediation, control/treatment, and prevention than have other industries.

    Figure 8: Electric generation industry expenditures (in $1000s) by year.

    Three non-generation electric industry entities, one a research consortium representing a number of public and private electric generating companies in the northeast, one a major public generation authority in the south, and one corporate office of a major private electric generation corporation in the southeast have spent $4,354,420 on the zebra mussel situation, 12.34% of all electric industry zebra mussel dollars (6.3% of all reported zebra mussel related expenses) since 1990. All three entities have been proactive responding to the zebra mussel issue, undertaking monitoring programs, rate payer education programs, prevention programs, and major R&D (research and development) programs.

    Twenty three hydro generating stations in 11 states responded to the survey, 13 (4,319,000 megawatts) of which are infested, and 22 (5,419,000 megawatts) of which spent money as a result of the mussels. The total hydropower expenditure was $1,759,000 (4.99% of electric industry expenses) with a mean expenditure of $79,950 per plant. Of this, $1,588,000 (90.28%) was spent at 13 infested facilities, $171,000 at nine noninfested sites (Figure 9). The largest hydro expense was $722,000 for chemical control (41% of all hydropower costs), with 10 facilities utilizing oxidizing chemicals and one using a nonoxidizing molluscicide (Figure 10). The mean chemical control expense per facility was $34,380. The second greatest expense ($393,000, 22.3%) was for planning and engineering expenses (mean expenditure of $17,865), followed by $255,000 (14.5%) for prevention programs (in all cases, this was for installation of systems for oxidizing chemical injection), and $222,000 (12.6%) for monitoring.

    Figure 9: Hydroelectric generation expenses (in $1000s) at infested and uninfested facilities, by year.


    Figure 10: Hydropower facility expenses by category (in $1000s).

    Eighty five fossil fuel powered electric generating stations in 19 states responded to the survey, 52 (34,970 megawatts) of which are infested, and 76 (48,539 megawatts) of which indicated that they had spent money as a result of zebra mussels. Sixty responses (70.6%) were from northern states, the remainder from southern. The total fossil fuel electric generating station expenditure was $11,067,200 (31.38% of total electric industry expenses, 16% of the total reported zebra mussel impact). Of this total, $10,868,200 (98.2%) was spent at the 52 infested facilities, with $199,000 spent at the 24 noninfested sites (Figure 11). The mean expenditure per fossil fuel station was $145,620.

    Figure 11: Fossil fuel electric generation facility expenses (in $1000s) at infested and uninfested facilities, by year.

    The largest fossil fuel plant expense was $3,670,500 for chemical control at 36 facilities representing 23,700 megawatts (33.17% of all fossil fuel costs, 10.41% of all reported power plant expenses), with 26 facilities utilizing oxidizing chemicals, 10 using molluscicides, and 10 using both oxidizing chemicals and molluscicides (Figure 12). The mean per plant chemical control expense was $101,960. The second greatest expense, $3,329,000 (30.08% of all fossil fuel costs, 9.44% of all power plant costs) was for retrofit projects at 20 facilities representing 10,323 megawatts (for chemical injection systems, cathodic protection systems, and waste heat recirculation replumbings), with a mean expenditure of $165,450. This was followed by $1,054,100 for monitoring expenses at 70 facilities (48,539 megawatts, a mean expenditure of $15,060, 9.52% of fossil fuel expenses, 3.0% of all power plant expenses). Other major expense categories were planning and engineering, research, and lost production ($475,000 at six facilities).

    Figure 12: Fossil fuel electric generation facility expenses by category (in $1000s).

    Responses were received from 23 nuclear power plants in 15 states and one province (representing a total generating capacity of 28,896 megawatts). Of these, 12 (15,877 megawatts) reported that they are infested by zebra mussels. All of the nuclear facilities reported spending money on zebra mussel related activities. Sixteen responses were from facilities located in the Great Lakes. Nuclear power plant expenditures totaled $18,093,400, 51.29% of the total electric generation industry impact (26.2% of the total reported zebra mussel economic impact), making this the largest water user category impact in the study. Of this, $17,607,900 (97.32% of nuclear expenses, 49.92% of the total electric industry cost) was spent at facilities that are infested by mussels, with the remaining $485,500 spent at uninfested facilities (Figure 13). The mean per facility expenditure was $786,670.

    Figure 13: Economic impact on infested and uninfested nuclear power plants (in $1000s) by year.

    The largest nuclear plant expense was $5,303,000 for facility retrofit at six plants with a total generating capacity of 6,209 megawatts (29.31% of all nuclear costs, 15.03% of all power plant expenses, 7.68% of all reported zebra mussel expenses). In each case, the retrofit was installation of source-end-of-pipe oxidizing chemical injection systems at an average of $883,800 per plant (Figure 14). The next largest category was $5,211,500 at 13 facilities (16,827 megawatts) for chemical control activities, an average of $400,900 per plant (28.8% of all nuclear expenses, 14.77% of all power plant costs, 7.55% of total reported expenditures). Five of the facilities reported using oxidizing chemicals, one uses molluscicides, and seven use both oxidizing chemicals and molluscicides. In addition, several other facilities also reported using chemical controls but did not quantify those expenses: two reported using oxidizing chemicals, one molluscicides, and three use both oxidizing chemicals and molluscicides. The third major expense category, $3,412,000 (18.86% of all nuclear plant costs, 9.67% of all power plant costs, 4.94% of all reported costs) was for prevention projects at three plants, in all cases chemical injection systems, at an average of $1,137,300 per plant. It should be noted here that these facilities considered the chemical control systems to be a prevention expense, and did not include it in their retrofit costs; it could be argued that these expenses were, in fact, a retrofit expense, which would increase retrofit to 48.17% of all nuclear facility expenditures (12.62% of all reported zebra mussel expenditures). Other major expenditures included: $1,580,000 at 10 facilities for planning and engineering expenses; $1,271,000 at five facilities for mechanical control activities; and, $968,900 for monitoring activities at 21 facilities.

    Figure 14: Nuclear generating station expenses (in $1000s) by category.