Biomass-Dependent Diet Shifts in Omnivorous Gizzard Shad: Implications for Growth, Food Web, and Ecosystem Effects

—We examined diet patterns of omnivorous gizzard shad Dorosoma cepedianum in Acton Lake, Ohio, during 1994–1997 using a multiple stable isotope analysis to quantify the role of this species in the system. On most dates, zooplankton were relatively depleted in d13C (about 230‰ to 225.5‰) compared with sediments (225‰), permitting construction of a mixing model to determine the proportion of C derived from benthic detritus and from planktonic productivity. During periods of greater gizzard shad biomass (.35 kg/ha), gizzard shad of more than 35-mm standard length (SL) derived most of their C from sediment detritus. When gizzard shad biomass was low (,15 kg/ha), zooplankton biomass increased and all sizes of gizzard shad derived most of their C from zooplankton. Conventional gut analyses corroborated these findings. Zooplanktivorous age-0 gizzard shad grew at three or more times the rate of those that were detritivorous. Rapid age-0 growth led to high gizzard shad biomass, a decrease in large zooplankton, and a subsequent shift to detritivory. Therefore, diet quality and growth rates are strongly linked to gizzard shad biomass, and these biomass-dependent feedbacks tend to keep gizzard shad biomass high in this system during most years. Because zooplanktivorous gizzard shad recycle nutrients within the water column, whereas detritivorous gizzard shad transport nutrients from sediments to the water column, biomass-induced diet shifts modify the impact of this species on phytoplankton through both top-down and bottom-up mechanisms. Knowledge of an organism’s diet is central to understanding its interactions with other species, including direct and indirect effects on food webs. Despite this importance and the prevalence of omnivory (Darnell 1961; Polis and Strong 1996), the diets of many omnivores are poorly understood, especially concerning the nutritional significance of detritus (Lemke and Bowen 1998). Detritivorous fishes and invertebrates are common in many aquatic communities, such as tropical rivers and streams (Bowen 1983), estuaries (Deegan et al. 1990), temperate streams (Cummins 1974), shallow lakes (Meijer et al. 1990), and reservoirs (Miranda 1984). Yet despite studies indicating that many species of fish ingest detritus in a facultative or obligate manner (Darnell 1961; Bowen 1983), little is known about its contribution to the nutrition of many omnivorous fishes (Ahlgren 1990; Lemke and Bowen 1998)—probably because of the difficulty in assessing its value to the longterm nutrition of an organism. Much of the energy flux and nutrient cycling within a variety of ecosystems occurs through the * Corresponding author: [email protected] 1 Present address: Department of Biology, Virginia Wesleyan College, 1584 Wesleyan Drive, Norfolk/Virginia Beach, Virginia 23502-5599, USA. Received May 24, 2000; accepted July 5, 2001 detritus food chain (e.g., Cummins 1974; Gosz et al. 1978; Polis and Hurd 1996; Polis and Strong 1996). Recent studies in aquatic systems have demonstrated that benthic-feeding omnivorous and detritivorous fish may have large impacts on ecosystem processes such as internal nutrient loading (Lamarra 1975; Brabrand et al. 1990), sediment resuspension (Meijer et al. 1990; Havens 1991) and maintaining high levels of fish productivity (Adams et al. 1983). The effects of omnivorous fishes on these processes depend highly on diet. For example, omnivorous fish feeding on benthic food sources (i.e., detritus, benthic invertebrates) may impact the benthos directly but also can transport nutrients into the water column via their excretions, serving as a net source of nutrients to pelagic phytoplankton, and potentially increasing total water column nutrients (Lamarra 1975; Shapiro and Carlson 1982; Persson 1997a). However, if these omnivores feed on plankton, they may have direct impacts through their planktivory but would not serve as a net source of nutrients to phytoplankton (Shapiro and Carlson 1982). This study examined the linkage between the diet and ecosystemic role of an omnivorous fish, the gizzard shad Dorosoma cepedianum, in a reservoir system to determine when this species functioned as a net source of nutrients to phytoplankton. By combining dietary data with information 41 DIET SHIFTS IN GIZZARD SHAD on nutrient excretion rates (Schaus et al. 1997) and their effects on phytoplankton and zooplankton (Schaus and Vanni 2000), we could assess the functional importance of this species in a reservoir ecosystem. To determine this value, we needed to quantify (1) contributions of benthic detritus versus zooplankton to the diet; (2) environmental or population characteristics that favor a particular mode of feeding; and (3) dietary consequences for the fish population (e.g., growth, reproduction). We used the gizzard shad as our model omnivore because it is abundant, has a wide distribution, and often dominates the fish biomass in Midwestern and southern reservoirs (e.g., Miranda 1984; Stein et al. 1995), especially those that are eutrophic (Bachmann et al. 1996; DiCenzo et al. 1996). Much is known about the food habits of gizzard shad. In some systems, adult gizzard shad consume zooplankton extensively (Drenner et al.1982; Mundahl 1988); in others, however, they feed extensively on organic detritus associated with sediments (e.g., Mundahl and Wissing 1987; Buynak and Mitchell 1993). Previous studies have demonstrated that gizzard shad can transport substantial quantities of nutrients into the water column when they feed on detritus (Schaus et al. 1997) and can have substantial effects on zooplankton (e.g., Drenner et al. 1982; Dettmers and Stein 1992; Schaus and Vanni 2000). Thus, diet quantification is critical to understanding the role of this species in food web and ecosystem processes. We used a multiple stable isotope technique to assess the contribution of detritus and zooplankton to the diet of gizzard shad. Stable isotope analyses have been used in food web studies to quantify carbon sources (e.g., Peterson et al. 1985; Keough et al. 1996), trophic structure (e.g., Kling et al. 1992; Hobson and Welch 1995), and migration between habitats (reviewed in Hobson 1999) to better quantify material transport and energy flow in food webs (e.g., Anderson and Polis 1998; Schindler and Lubetkin, in press). They also have been successfully used to quantify the importance of detritus to the diets of omnivorous fishes in other systems (Araujo-Lima et al. 1986; Deegan et al. 1990; Forsberg et al. 1993). This technique integrates feeding history over the time scale of tissue turnover and quantifies the importance of potential food sources (Peterson and Fry 1987). Whereas conventional gut samples provide only a ‘‘snapshot’’ of items in the gut over a very short time interval, isotopes measure actual assimilation of various food sources and thus account for selective digestion of particular food items (Peterson and Fry 1987). In this study, we coupled isotopic analyses with conventional gut analyses, estimation of gizzard shad biomass, and environmental sampling (zooplankton and seston abundance) to predict the conditions that facilitate a particular mode of feeding.