Disentangling the ecological role of gizzard shad in Kansas small impoundment food webs : changes in limnology, fish communities, and predator diets after rotenone manipulation

Abstract

Benthic-feeding fishes can exert direct and indirect effects on multiple trophic levels in lentic ecosystems by increasing water column nutrient concentrations, stimulating phytoplankton abundances, and increasing turbidity. One benthic-feeding fish that has been the target of selective biomanipulations is the gizzard shad (Dorosoma cepedianum), a freshwater clupeid. In reservoirs across the United States, gizzard shad often dominate fish assemblages in terms of relative abundances and proportions of total fish biomass, especially in eutrophic to hypereutrophic systems. I tested the influence of gizzard shad on lentic ecosystems by sampling key limnological parameters in 12 small impoundments across a range of gizzard shad relative abundances over four years. I hypothesized that gizzard shad relative abundance would be positively associated with algal abundance, cyanobacterial abundance, and nutrient concentrations. Higher gizzard shad relative abundances were associated with elevated total nitrogen (TN) and total phosphorus (TP) during certain months and with a general reduction in zooplankton densities. Regardless of the apparent influence of gizzard shad abundance on top-down (zooplankton) and bottom-up (nutrient) processes that might affect primary producers, I found no relationship between gizzard shad relative abundance and cyanobacteria or total phytoplankton abundance. In a manipulative experiment, I tested whether selectively eradicating gizzard shad from food webs via replicated low-dose rotenone applications could precipitate responses in abiotic and biotic variables. I collected baseline measurements of limnological and fish community parameters from eight impoundments with gizzard shad for two years (2017 and 2018) before rotenone was applied to four treatment lakes in December 2018 and then monitored food web components for two years posttreatment (2019 and 2020). TN and TP decreased in treatment impoundments relative to control impoundments following shad removal, but I found little evidence that gizzard shad removal had a substantial impact on limnological properties or fish populations. Finally, because gizzard shad are thought to be important forage for piscivorous fishes in reservoirs, I characterized the importance of gizzard shad in largemouth bass diets in 12 Kansas small impoundments and examined how piscivory varied across bass length classes, months, and a gradient of gizzard shad relative abundances, hypothesizing that largemouth bass would selectively prey upon gizzard shad compared to other forage fishes. Bass length class was a significant factor influencing the frequency of occurrence of fish (piscivory) in bass diets in our regression model, but piscivory was not associated with sampling month or gizzard shad relative abundance. Gizzard shad eradication did not alter the frequency of occurrence of fish in largemouth bass diets following rotenone treatment. Since the contribution of gizzard shad prey to bass diets was low relative to the contribution of centrarchids, cannibalism, and crayfish, selective shad removal via rotenone biomanipulation might not meaningfully alter largemouth bass diet composition. Overall, gizzard shad effects were weak and mixed, and our ability to detect responses was greatly limited by our small sample size of impoundments. My dissertation work sheds light on the ecological role of gizzard shad in small impoundments and illustrates the complexity of studying highly stochastic novel ecosystems managed for angling recreation.