USGS - science for a changing world

U.S. Geological Survey - Great Lakes Science Center

Prey Fish Assessment

Since the 1970s, the USGS-GLSC has collected long-term data on prey fish communities in all five Great Lakes. The prey fish assemblage, which includes alewife, gizzard shad, emerald shiner, rainbow smelt, bloater, sculpins, and lake herring, is a vital trophic link in Great Lakes ecosystems connecting aquatic invertebrates and plants to predator fish. USGS-GLSC researchers use tried-and-true methods, such as bottom and mid-water trawling, as well as advanced technologies, such as hydroacoustics, to assess preyfish populations. The dataset generated through this work is critical to helping resource managers determine accurate stocking levels for predator fish in the Great Lakes, and is also one of the most highly valued sources for understanding human influences in the Great Lakes, guiding native fish restoration, and assessing invasive species impacts on fisheries production.


Lake Erie’s western basin harbors the most productive fish community in all of the Great Lakes. The western basin is also one of the most significant contemporary and historical nursery areas for fish populations in Lake Erie. LEBS scientists routinely survey the fish community in the western basin to understand growth, recruitment, and diets of key predators (such as yellow perch, white perch, and walleye), as well as the abundance, distribution, and habitat use of many different fish populations. The information developed from this monitoring work is utilized by Lake Erie Task Groups for improved inferences about the status and trends of fish populations and the overall health of the western basin ecosystem.

Principal Investigator: Richard Kraus

Bottom trawl sampling at our East Harbor sites has been conducted since 1961. This long-term data set has permitted understanding how recruitment indices of native benthic species have been affected by invasive species. Data have also been used to inform species recovery plans for Silver Chub, a COSEWIC species of special concern. Ongoing work is examining ecology of Silver Chub and Trout-perch, the former a species recovering from very low numbers whose recovery coincided with establishment of dreissenid mussels, the latter an abundant benthic species with no known predators.

Principal Investigator: Patrick Kocovsky

Hydroacoustics permits sampling large areas over short periods of time with minimal negative effects on aquatic fauna and has been used to develop indices of abundance of several Great Lakes forage species (e.g., emerald shiner, rainbow smelt) for over 2 decades. Continual improvement of hydroacoustics methods through research into data collection and analysis techniques has been part of GLSC hydroacoustics programs for the past decade. Hydroacoustics has been identified by the CLC and Lake Erie partners as a high-priority program for the GLSC. Current collaborative research includes: evaluating elements of the Great Lakes Standard Operating Procedure and sensitivity analysis of parameters for analyzing hydroacoustic data.

Principal Investigator: Patrick Kocovsky

Since it became a lakewide survey in 1973 (i.e., sampling 7 transects consistently: Manistique, MI; Frankfort, MI; Ludington, MI; Saugatuck, MI; Waukegan, IL; Port Washington, WI; Sturgeon Bay, WI) the Lake Michigan bottom trawl study has played a critical role in understanding the ecosystem dynamics and in managing the fisheries of Lake Michigan. Its primary role is to provide annual estimates of prey fish abundance to guide the decision of state agencies in the stocking of piscivorous fish. Beyond these data, GLSC researchers have analyzed the time series to 1) further understanding of the factors that regulate the recruitment of alewife, bloater, deepwater sculpin, and slimy sculpin, 2) determine trends in the growth, maturation, and condition of alewife and bloater, and 3) conduct community analyses to explore how the prey fish community has responded to invasive species, such as alewife and dreissenid mussels. Maintenance of this long-term data set is critical to provide a benchmark for how these important fishes are responding to the persistent perturbations to the Lake Michigan ecosystem, include invasive species and climate change.

Principal Investigator: David "Bo" Bunnell

Lake Superior historically supported a diverse assemblage of native coregonid species that co-evolved to utilize available habitats (Koelz 1929).  Many of these forms thrive to the present day (Zimmerman and Krueger 2009), but shortjaw cisco (Coregonus zenithicus) are currently being considered for listing under the Endangered Species Act (ESA). Unfortunately, there exists a great deal of variation in body shapes of Coregonus spp. in the lake which complicates species identification.  In a recently funded SISNAR project, we compiled data on body measurements of three morphotypes of cisco (Coregonus artedi) provided by Koelz (1929) and developed a discriminant function (DF) model to assign contemporary cisco to morphotype based on photographs (Yule et al. in press). Our analyses showed that the historic variation in cisco body shapes that existed prior to their collapse in the lower Great Lakes can still be found in the basin.  Since that time, we have had opportunity to enter and analyze Koelz (1929) data on the entire Lake Superior coregonid flock. We have developed a DF model based on body measurements, and gillraker counts and measurements that can identify coregonid species identical to Koelz (1929) with nearly 100% accuracy. 

During 2011-2012, USGS-Great Lakes Science Center partnered with the Ontario Ministry of Natural Resources to gather upwards of 2500 digital photographs of coregonids, saving the front half of their bodies so we could recover their gillrakers.  With this proposal, we seek funding to hire a student intern over the summer to analyze these collections.  We have 2 objectives: 1) determine if the two collecting agencies are identifying coregonid species similar to Koelz (1929), and 2) determine the degree that status information of each species varies depending on the field identifications and the species assignments developed with our DF model.  This project is relevant to native people because many band members hold fishing rights that could be negatively impacted depending on decisions levied by ESA.  Since the study of Koelz (1929) anchors our understanding of the status of current populations, it is critical that we demonstrate that coregonid species are being identified as they were in the past.

Principal Investigator: Dan Yule

This study had two components, 1) researching early life history dynamics, and 2) modeling the stock-recruit relationship for Lake Huron rainbow smelt.  This study is in the finishing stages with one of two planned manuscripts published and the second manuscript near completion.

Early life history dynamics of rainbow smelt - The specific research objectives for this study were:

  1. Determine the influence of tributary water temperatures on the timing and duration of rainbow smelt spawning during April.
  2. Determine the relative abundances, growth rates, and mortality rates of larval rainbow smelt cohorts in St.Martin Bay, Lake Huron and relate these population dynamics to bay water temperatures.

To understand the influence of water temperature on reproduction, growth, and survival during larval-fish stages, we sampled spawning tributaries and larval-fish habitats during 2008 and 2009 in St. Martin Bay, Lake Huron.  Spawning by rainbow smelt occurred primarily when stream temperatures were between 3 and 10 oC, which resulted in a 7-10-day spawning period during 2008, and a 15-20-day spawning period in 2009.  Regardless of these differences in spawning temperatures and duration, peak larval-fish densities during 2008 were double those observed during 2009.  Length-frequency analysis of larval-fish populations during both years revealed stream-hatched fish during May and a later emergence of larval rainbow smelt during summer, presumably originating from lake spawning.  Warmer bay water temperatures led to earlier emergence of lake-spawned rainbow smelt larvae during 2009.  Stream-hatched fish larvae experienced large-scale mortality during May 2008 resulting in a bay population consisting primarily of lake-spawned rainbow smelt larvae, but during 2009 both stream- and lake-hatched cohorts experienced higher survival concomitant with significantly higher mean population growth rates.  Higher larval-fish growth rates during 2009 appeared to be density-dependent and facilitated by warmer water temperatures during late June and cooler water temperatures during July.  Temperature-mediated differences in annual growth rates and irregular contributions from stream- and lake-hatched fish larvae are important factors affecting survival and abundance of young-of-the-year rainbow smelt in Lake Huron.

Stock-recruit modeling - Our goal was to evaluate key ecological factors influencing age-0 Rainbow Smelt recruitment and to examine the relationship between predators and age-1 and older Rainbow Smelt abundance in Lake Huron.  Our specific research objectives were:

  1. Determine the influence of Rainbow Smelt stock biomass on the production of age-0 recruits.
  2. Determine the influence of selected biotic (i.e., Lake Trout and Chinook Salmon abundance) and abiotic (i.e., water temperature, lake levels, and precipitation) ecological factors on recruitment.

Rainbow Smelt Osmerus mordax are native to Northeast Atlantic and Pacific-Arctic drainages and have been widely introduced or dispersed in North America.  Rainbow Smelt are known predators and competitors of native fish and a primary prey species in the Great Lakes region.  Despite their widespread distribution, importance as a prey species, and potential for negative interactions with native fish species, there is a paucity of information concerning recruitment dynamics for Rainbow Smelt.  We constructed Ricker stock-recruit models for Rainbow Smelt in Lake Huron based on bottom trawl catches to determine the influence of stock size, environmental factors (water temperature, lake levels, and precipitation) and salmonid predation on the production of age-0 recruits from 1976-2010.  Rainbow Smelt recruitment was negatively related to stock size indicating that compensatory, density-dependent mortality from cannibalism or intraspecific competition was an important factor related to the production of age-0 recruits.  Recruitment was positively related to spring precipitation suggesting stream spawning habitats were modified by precipitation to provide suitable conditions for egg and larval Rainbow Smelt survival.  Rainbow Smelt recruitment was positively related to Lake Trout abundance.  However, spawning stock biomass of Rainbow Smelt, which declined substantially from 1989-2010, was negatively associated with Lake Trout CPE suggesting predation was an important factor related to the decline of age-1 and older Rainbow Smelt in Lake Huron.  Recruitment of Rainbow Smelt in Lake Huron is regulated in part by competition with or cannibalism by older conspecifics, spring precipitation, and predation by Lake Trout on age-1 and older Rainbow Smelt.

Principal Investigator: Tim O'Brien

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo logo U.S. Department of the Interior | U.S. Geological Survey
Page Contact Information: GLSC Webmaster
Page Last Modified: Thursday December 15, 2016