Herring and other small pelagic fishes represent a trophic bottleneck for the fluxes of energy and biomass from marine primary producers up to the level of apex predators (Checkley et al., 2009)—including humans (Konar et al., 2019). Atlantic herring (Clupea harengus) itself is not only an important prey for numerous marine predators (Overholtz & Link, 2006) but also one of the most caught marine fish species worldwide (2020 landings: 1.6 million metric tons; FAO, 2022). The species' importance led to an outstanding research effort during the last two centuries making Atlantic herring one of the best studied fish species worldwide (Geffen, 2009). According to the plethora of studies conducted on herring, the species is a commonly accepted zooplankton feeder (e.g., Bachiller et al., 2016), at least at its offshore feeding ground, where almost all of the investigations had been conducted (Möllmann et al., 2004). However, most subpopulations of Atlantic herring migrate seasonally between their offshore habitats and coastal waters (Miethe et al., 2013), but their feeding traits in coastal systems have been less extensively studied and are often assumed to resemble their behavior at the open sea (Elliott et al., 2002). Such a transfer of feeding behavior patterns to other systems and seasons is common practice when parameterizing multispecies population models for fisheries management but always carries the risk of introducing uncertainties into the models. In 2010, we started examining the stomach contents of herring during their spring and fall migrations into Greifswald Bay, a shallow lagoon in the Baltic Sea (54.2° N; 013.5° E; Appendix S1: Figure S1) and a crucial spawning ground for herring. Our primary interest was to study the potential effects of filial cannibalism (the act of preying on offspring of its own species) on the survival of herring larvae. Surprisingly, adult herring caught in fall were exceptionally filled with prey (up to a maximum of 12.3% of the total body weight) but with neither the expected mesozooplankton nor conspecific larvae but with demersal fishes (mainly gobies of the genus Pomatoschistus) and large invertebrates (Figures 1 and 2). Intrigued by this observation, we repeated the sampling in two more successive years (Appendix S1: Table S1) and found the same feeding behavior in 2011 and 2012 (Figure 2) and during later sampling campaigns (2014, 2015, and 2020; Appendix S1: Table S2). In the stomach of a single herring sampled in fall 2012, we found at least 31 gobies between 11 and 43 mm total length. Another individual had consumed three smelts (Osmerus eperlanus) with individual total lengths between 53 and 84 mm (Figure 1). While many stomachs contained a high proportion of benthic invertebrates in fall (up to a maximum of 1478 mg per fish; mainly shrimps, amphipods, and mysids), copepods formed only a minor part of these stomach contents (Figure 2), although we found high abundances of this typical herring prey in the water (between approximately 21,000 and 95,000 copepods/m2 in spring 2011 and 21,000–28,000 copepods/m2 in fall 2011, which is comparable to the numbers usually found in offshore herring feeding areas; Appendix S1: Tables S3 and S4). Consequently, a limitation in copepods cannot explain the observed feeding behavior. It is more likely that the better nutritional value of prey species from higher trophic levels (Zheng et al., 2021) might lead to opportunistic predation on fish (e.g., larvae), which would be of advantage in the subsequent, food-limited winter. However, we can only speculate why fall stomachs contained no herring larvae at all (Figure 2) but the relatively low larvae abundances at this time of year (compared to most weeks sampled in spring, Appendix S1: Table S5) may have made them a less attractive prey. In contrast to the fall stomachs, most spring stomachs were empty (76%, Figures 1 and 2). Even the filled spring stomachs did not exceed 3% of the stomach content weight found in fall, resulting in highly significant seasonal differences (Kruskal–Wallis χ2 = 187.08, degrees of freedom (df) = 5, p-value < 0.001, Appendix S1: Table S6). On a horizontal scale, herring adults and larvae overlapped well during spring season (Appendix S1: Table S5), and findings by Polte et al. (2017) led to our assumption that both groups also overlapped well within the vertical water column. Considering the numerous studies and reports on the potential effects of filial cannibalism and intraguild predation on the population dynamics of small pelagic fishes (e.g., Bachiller et al., 2015; Irigoien & de Roos, 2011; Szeinfeld, 1991), we expected to find adult herring feeding on the larvae. However, we found in merely four of all 713 examined stomachs a single herring larva each (Figure 2), and seasonal differences in herring larvae abundances (Appendix S1: Table S7) were not reflected by corresponding changes in the larvae contribution to the stomach contents. This suggests that filial cannibalism is not a structuring factor in early life stage survival of this herring subpopulation. Refusing food ingestion during the reproductive season can be observed in many different species of fish (McBride et al., 2015) and is probably linked to physical changes during the maturation process. Whether this is merely a physiological side effect or not, such a feeding behavior results in reduced offspring mortality for this species. Our observations further demonstrate additional mechanisms of a coupling between benthic and pelagic food webs. These systems are usually considered separately in oceanic environments, but in the shallow coastal waters the boundaries between them become blurred in the limited space between the surface and the seabed. Immigrating oceanic species might therefore trophically interlink in unusual ways, thereby promoting coupling processes between distinct water bodies and ecosystems. The phenomenon of marine fish migrating toward coastal and riverine systems for spawning (e.g., salmon, herring) has so far mainly been related to a massive import of biomass and energy (Fox et al., 2018). For example, many different coastal predators prey upon adult herring and their eggs and larvae (Kotterba et al., 2017; Skaret et al., 2015). Our findings underpin that inverse fluxes do also exist (in our case, in the shape of goby-eating, ocean-going herring), which need to be considered in future models that aim to characterize the coupling between oceanic and coastal systems. Integrating trophodynamic interactions into the population models for exploited wildlife such as marine fish stocks is appealing for management as this can help to improve the prediction range and accuracy of the models (Pikitch et al., 2004; Trijoulet et al., 2019). Although fisheries biologists and managers worldwide strive to include environmental data in their assessment, only approximately 2% of regulated global fish stocks are currently managed using such ecosystem-based approaches (Skern-Mauritzen et al., 2016). Feeding traits of marine fish are often disproportionally investigated (Funk et al., 2021), and the corresponding knowledge gaps that hinder a parameterization of the models are usually filled by transferring known patterns across species, areas, seasons, and life stages (e.g., Sánchez-Hernández et al., 2019). However, the reliability of models based on such assumptions strongly depends on the level of feeding opportunism and spatiotemporal stability (Funk et al., 2021). Although copepods probably remain a major prey for herring throughout the year, the general capability of the species to seasonally feed on higher trophic levels such as demersal fish should be considered in future bioenergetic studies. According to our observations, the positioning of herring within the estuarine food web should be revised by shifting it to a higher trophic level (Figure 3). Although herring predation on larval fish (Hallfredsson & Pedersen, 2009), benthic invertebrates (Arrhenius & Hansson, 1993), and, occasionally, small fishes (Hessle, 1925; Köster & Möllmann, 2000) has been reported, our study showed new evidence for a stable and annually recurrent seasonal feeding pattern. Integrating such feeding traits into future models will require more holistic approaches in food web science that consider all relevant seasons, habitats, and the spatiotemporal distribution of noncommercial but crucial species. This work was supported by the European Union and the Federal Ministry of Education and Research of Germany (BONUS BIO-C3, Grant 03F0682B; SpaCeParti, Grant 03F0914E; and balt_ADAPT, Grant 03F0863D) and the Data Collection Framework (DCF) of the European Union. We thank our colleagues at the Thünen-Institute of Baltic Sea Fisheries and aboard the research vessel Clupea for their great support. We further thank Myron Peck, Matthias Paulsen, Cornelius Hammer, and Daniel Oesterwind for their comments on a very early manuscript version. Open Access funding enabled and organized by Projekt DEAL. The authors declare no conflicts of interest. Data (Kotterba, 2024) are available in Zenodo at https://doi.org/10.5281/zenodo.11187517. Appendix S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
No takes yet. Share an insight, caveat, or question.
Kotterba et al. (2024) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: