Elsevier

Fisheries Research

Volume 236, April 2021, 105842
Fisheries Research

Modeling round goby growth in Lake Michigan and Lake Huron with multi-model inference

https://doi.org/10.1016/j.fishres.2020.105842Get rights and content

Abstract

Although the round goby Neogobius melanostomus has become established throughout the Laurentian Great Lakes, a multi-model inference (MMI) approach toward characterizing round goby growth in the Laurentian Great Lakes has yet to applied using otolith-derived data. Further, spatial variation in round goby growth among lakes has yet to be investigated. For each sex, growth of round gobies at three locations of Lake Michigan and four locations of Lake Huron was investigated using MMI, based on information theory, with three candidate growth models. These three growth models included the von Bertalanffy model, the Gompertz model, and the logistic model. The von Bertalanffy model was most often selected (13 out of 14 cases) as the ‘best’ model among all candidate models, followed by the logistic model. None of the best models were strongly supported as a ‘clear winner’. At least one additional model was supported by the data in each of the 14 cases, indicating that there is a substantial degree of uncertainty in model selection. When model selection uncertainty was ignored, standard errors of growth parameters were underestimated in 8 of the 14 cases. Overall, round gobies in Lake Michigan attained larger sizes at age and grew faster than in Lake Huron. Based on multi-model inference, our study provided a robust assessment of round goby growth, which will be essential in better managing sport fisheries in both lakes.

Introduction

The round goby Neogobius melanostomus, a small, benthic fish native to the Ponto-Caspian region, is one of the most successful invaders to have widely extended its distribution range both in Europe and North America (Kornis et al., 2012). This species is considered to be transferred and introduced to new locations by means of ship ballast-water transport (Claudi and Ravishankar, 2006). In North America, the round goby was first observed in the St. Clair River in 1990 (Jude et al., 1992). Since that time, the round goby has become well established in the Laurentian Great Lakes and has become one of the most ecologically significant aquatic invasive species in the Great Lakes basin (Johnson et al., 2005; Kornis et al., 2012; He et al., 2015). Predicted ecological impacts induced by round goby colonization include predation on eggs and fry of native fishes (Chotkowski and Marsden, 1999; Phillips et al., 2003; Steinhart et al., 2004; Roseman et al., 2006), exclusion of native benthic fish species due to competition for spawning space, habitat, and dietary resources (Jude et al., 1995; Jude, 1997; French and Jude, 2001; Janssen and Jude, 2001; Bergstrom and Mensinger, 2009), service as a conduit to funnel dreissenid energy production to upper trophic levels of the food web (Johnson et al., 2005; Campbell et al., 2009), and support of piscivore biomass by providing an alternative energy source (Truemper and Lauer, 2005; Dietrich et al., 2006; Jacobs et al., 2010; Madenjian et al., 2011; Pothoven and Madenjian, 2013; He et al., 2015; Happel et al., 2018; Mumby et al., 2018). Therefore, it is clear that round gobies are becoming increasingly important members of Great Lakes food webs.

Successful invaders often display diversity in life-history characteristics. Understanding life-history characteristics, such as growth and age structure, is vital for evaluating long-term effects of invaders on food webs and for developing practical management techniques for the control of invaders (Veitch and Clout, 2002). Previous studies on round goby age and growth within invaded ecosystems indicated that size within an age group was highly variable and males grew faster and attained a larger maximum size than females (MacInnis and Corkum, 2000a, 2000b; Phillips et al., 2003; French and Black, 2009; Gümüs and Kurt, 2009; Taraborelli et al., 2010; Sokołowska and Fey, 2011; Huo et al., 2014; Duan et al., 2016). Furthermore, life-history characteristics were found to vary over time and space. With regard to time, Gruľa et al. (2012) reported that adult round gobies in the middle Danube River allocated energy resources to reproduction and offspring care in newly established populations, rather than somatic growth, when the population was newly established. However, after the population had been established for some time, there was a shift in allocation towards increased somatic growth. Lynch and Mensinger (2013) determined that round gobies in Duluth Harbor of Lake Superior demonstrated maximum growth in July and August and almost no growth between October and March. With regard to space, Huo et al. (2014) reported that round goby growth showed significant spatial variation within Lake Michigan. Round gobies at Sleeping Bear Dunes and Waukegan grew significantly faster than those at Sturgeon Bay. Similarly, round gobies exhibited significant spatial variation in growth within Lake Huron (Duan et al., 2016). Moreover, round gobies in tributaries to Lake Michigan grew much faster and appeared to have shorter life spans than round gobies residing in Lake Michigan proper (Kornis et al., 2017). Although several investigations have been conducted on age and growth of round gobies in invaded aquatic ecosystems, none have directly focused on spatial variation in round goby growth between lakes.

Robust characterization of round goby growth within each of the Laurentian Great Lakes may be useful in better managing sport fisheries in each of these lakes. At present, a multispecies, age-structured predator-prey model is being used to guide decision making for managing sport fisheries in Lake Michigan (Tsehaye et al., 2014; Luo et al., 2019), and development of a similar predator-prey model is underway for Lake Huron (J. Bence, Michigan State University, personal communication). Prey fish growth is an integral component of these predator-prey models (Tsehaye et al., 2014). With the recent increased importance of round gobies in the diet of these sport fish, our assessment of round goby growth in both of these lakes will be critical in refining these predator-prey models to better guide management of the sport fisheries. Note that the abovementioned predator-prey models have been developed at a lakewide level of scale, because the predators are highly mobile and may travel throughout the lake. Thus, growth of the prey needs to also be summarized at a lakewide level so that it is compatible with the lakewide scale of the modeling approach.

A mathematical expression of the mean individual body growth is fundamental for population analysis, stock assessment, and fisheries management. Several models have been proposed to estimate the mean individual growth in a population. The von Bertalanffy growth model is the commonly applied model in length-age data analysis (von Bertalanffy, 1938). However, other models like the Gompertz model (Gompertz, 1825) or the logistic model (Ricker, 1975) have been shown to better describe absolute growth for many aquatic species. “What is the best model to use?” is the most often thought and the critical question in making valid inference from the entire set of models. Two fundamentally different approaches are frequently used to answer this question. One is to fit by the ‘selected’ model, in which inference and estimation of parameters and their precision are limited to those of the selected model. Another approach is to fit more than one model and then use a criterion to select the ‘best’ model. However, the ‘best’ model, when used for inference without actually being the best model, could lead to biased point estimation and false evaluation of precision of growth parameters, and this ‘best’ model selection approach could not be used to quantify the relative support for the various models (Johnson and Omland, 2004; Katsanevakis and Maravelias, 2008).

Model selection based on information theory is a relatively new paradigm in biological sciences and has been recommended as a better and more robust alternative than traditional approaches (Burnham and Anderson, 2002). The information theory approach frees the researcher from the limiting concept that the proper approximating growth model is somehow ‘given’. When data support evidence of more than one model, model averaging the predicted response variable across models is advantageous in reaching a robust inference that is not conditional on a single model. Rather than estimating parameters from only the ‘best’ growth model, parameter estimation can be made from several or even all the models considered. This procedure is termed multi-model inference (MMI) and has several theoretical and practical advantages (Burnham and Anderson, 2002; Katsanevakis, 2006; Rabaoui et al., 2007; Katsanevakis and Maravelias, 2008).

Despite the increasing importance of round gobies in food webs of lakes where this species has invaded, MMI has yet to be used to characterize round goby growth. Moreover, round goby growth has typically not been rigorously compared and contrasted across lakes. The primary goal of our study was to model round goby growth with MMI to make robust parameter estimations, and the secondary goal of our study was to determine whether round goby growth showed significant spatial variation between Lake Michigan and Lake Huron. Robust characterization of round goby growth in both Lake Michigan and Lake Huron can be applied toward better management of the sport fisheries in both lakes. If round goby growth significantly differed between the two lakes, then the round goby growth model for Lake Michigan would not be appropriate for use in the Lake Huron predator-prey model and vice versa.

Section snippets

Data sets

In total, seven sets of total length-age data for both sexes were taken from the literatures for round gobies: three datasets from Lake Michigan, including the Waukegan, Sturgeon Bay, and Sleeping Bear Dunes locations (Huo et al., 2014), and four datasets from Lake Huron, including the Saginaw Bay, Rockport, Hammond Bay, and Thunder Bay locations (Duan et al., 2016). Round gobies were collected in Lake Huron at depths between 3 and 27 m using bottom trawls and minnow traps in 2014, whereas fish

Length composition and age structure

The size distribution of round gobies was shown in Fig. 1 for all seven locations. Round gobies collected in Lake Michigan were larger in total length than in Lake Huron for both sexes. In Lake Michigan, round gobies ranged in total length from 49 to 131 mm for males and from 48 to 119 mm for females, respectively, while in Lake Huron total length ranged from 44 to 123 mm for males and from 45 to 115 mm for females, respectively.

Age estimates ranged from 2 to 7 years for both sexes in Lake

A robust approach for round goby growth estimation

We found that none of the best models were strongly supported as a ‘clear winner’ (wi > 0.9). At least one additional model was supported by the data in each case. When the data support more than one model, there is a substantial degree of uncertainty in model selection that cannot be ignored. Generally, model selection uncertainty causes a magnification of the standard error of the parameters of the best model (Katsanevakis and Maravelias, 2008). Our results indicated that underestimation of

CRediT authorship contribution statement

Youjian Duan: Writing - original draft, Formal analysis, Methodology, Software, Visualization. Charles P. Madenjian: Writing - review & editing, Funding acquisition, Resources, Supervision, Validation. Yingming Zhao: Writing - review & editing, Validation. Bin Huo: Conceptualization, Writing - review & editing, Methodology, Funding acquisition, Visualization.

Declaration of Competing Interest

The authors report no declarations of interest.

Acknowledgments

We thank Andrew Briggs of the U. S. Fish and Wildlife Service Detroit River International Wildlife Refuge for providing round gobies from Rockport, Todd Wills and Mike Thomas of the Michigan Department of Natural Resources Lake St. Clair Fisheries Research Station for providing round gobies from Saginaw Bay, Henry Thompson and Zach Wickert of the USGS Hammond Bay Biological Station for collection of round gobies in Hammond Bay; Patrick Hudson and Kevin Keeler for assisting with the use of the

References (52)

  • T.B. Johnson et al.

    A potential new energy pathway in central Lake Erie: the round goby connection

    J. Great Lakes Res.

    (2005)
  • S. Katsanevakis

    Modeling fish growth: model selection, multi-model inference and model selection uncertainty

    Fish. Res.

    (2006)
  • J.A. Mumby et al.

    Diet and trophic niche space and overlap of lake ontario salmonid species using stable isotopes and stomach contents

    J. Great Lakes Res.

    (2018)
  • E.C. Phillips et al.

    The round goby (Neogobius melanostomus) in Pennsylvania tributary streams of Lake Erie

    J. Great Lakes Res.

    (2003)
  • E.F. Roseman et al.

    Predation on walleye eggs by fish on reefs in western Lake Erie

    J. Great Lakes Res.

    (2006)
  • A.C. Taraborelli et al.

    Round goby (Neogobius melanostomus) population structure, biomass, prey consumption and mortality from predation in the Bay of Quinte, Lake Ontario

    J. Great Lakes Res.

    (2010)
  • M.A. Bergstrom et al.

    Interspecific resource competition between the invasive round goby and three native species: logperch, slimy sculpin, and spoonhead sculpin

    Trans. Am. Fish. Soc.

    (2009)
  • D.B. Bunnell et al.

    Changing ecosystem dynamics in the Laurentian Great Lakes: bottom-up and top-down regulation

    BioSci.

    (2014)
  • K.P. Burnham et al.

    Model Selection and Multi-model Inference: a Practical Information-Theoretic Approach

    (2002)
  • Y. Chen et al.

    A comparison of von Bertalanffy and polynomial functions in modelling fish growth data

    Can. J. Fish. Aquat. Sci.

    (1992)
  • R. Claudi et al.

    Quantification of risks of alien species introductions associated with ballast water discharge in the Gulf of St

    Lawrence. Biol. Invasions

    (2006)
  • J.R.P. French et al.

    Maximum length and age of round gobies (Neogobius melanostomus) in Lake Huron

    J. Freshw. Ecol.

    (2009)
  • B. Gompertz

    On the nature of the function expressive of the law of human mortality and on a new mode of determining the value of life contingencies

    Philos. Trans. R. Soc. Lond.

    (1825)
  • D. Gruľa et al.

    Age and growth of invasive round goby Neogobius melanostomus from middle Danube

    Cent. Eur. J. Biol.

    (2012)
  • A. Gümüs et al.

    Age structure and growth by otolith interpretation of Neogobius melanostomus (Gobiidae) from southern Black Sea

    Cybium.

    (2009)
  • A. Happel et al.

    Spatial variability of lake trout diets in Lakes Huron and Michigan revealed by stomach content and fatty acid profiles

    Can. J. Fish. Aquat. Sci.

    (2018)
  • Cited by (2)

    View full text