Forage fish for cod and people

Ecosystem-based fishery management (EBFM), which is now widely accepted (1–3), calls for consideration of ecological linkages between marine species when regulating harvest levels. However, whether and how to alter harvest rates and minimum biomass limits (relative to conventional single-species approaches) to protect ecosystem integrity is less settled. For forage fish species—small fish such as herring, capelin, anchovies, and sardines that are important prey for higher trophic level species in marine ecosystems—there are calls for more precautionary harvest strategies than have typically been employed. Smith et al. (4) recommend halving exploitation rates, and Pikitch et al. (5) call for reducing harvest rates progressively toward zero at 40% of unfished biomass for stocks with intermediate information, and maintaining forage stocks at over 80% of unfished biomass when information is low. However, there is little empirical evidence that predator populations have been impacted by fishery-induced swings in forage fish populations, and trophic modeling studies have tended to overestimate the impact of forage species on predators (6). The analysis of Soudijn et al. (7) in PNAS suggests that curtailing or closing down forage fish fisheries could have unintended negative consequences on some predators. Their analysis suggests that moderate and even high levels of exploitation of forage fish stocks may actually help prevent the collapse of piscivorous (fish-eating) predators such as cod that are subject to heavy fishing pressure.

Soudijn et al. (7) use a community dynamics model of Baltic sea cod and key forage fish stocks (herring and sprat) to evaluate how exploitation rates of the forage species impact productivity of cod as exploitation rates for cod vary. They complement their community dynamics modeling analysis with a broad statistical analysis of 23 overlapping combinations of forage fish and their predators from around globe. They evaluate correlation of declines in piscivorous fish stocks with exploitation levels of both the piscivores themselves and forage stocks in the same region. Both analyses support the hypothesis that relatively high fishing mortality on forage stocks may actually benefit piscivore stocks that are also undergoing high fishing mortality and are in decline. The rationale for this result, which is modeled directly with the community dynamics model, is that fishing down the forage population reduces intraspecific competition, enabling greater reproduction of juveniles, which are key forage for the younger piscivores cohorts that are critical to maintaining or rebuilding the heavily fished piscivore stock. The statistical analysis by Soudijn et al. (7) is correlative, so causation and the mechanisms driving the results are uncertain, but the results are consistent with the hypothesis that higher exploitation rates of forage fish reduce declines of piscivores that are also heavily fished.

The findings of Soudijn et al. (7) complement other studies that raise questions about whether higher forage fish abundance will necessarily increase growth or inhibit declines of predators. Walters and Kitchell (8) argue that forage fish stocks that increase as a result of declines in their predators may undermine recovery of piscivores if they prey on or compete for resources with juveniles of the piscivore population. Köster and Möllmann (9) find herring and sprat represent a substantial source of egg mortality for cod in the Baltic, although they note that this does not necessarily mean they are limiting cod recruitment. This predation effect could strengthen the effect of intraspecific competition modeled by Soudijn et al. (7). Collie et al. (10) show that predation by herring on cod eggs and larvae could delay rebuilding of Georges Bank cod, and Essington et al. (11) find a similar result with a multispecies model of North Sea cod and herring. These are modeling studies that demonstrate how the mechanisms might work rather than prove they have done so, but empirical analyses also raise doubts that piscivore productivity is directly related to abundance of individual forage fish stocks. In a study of North American ecosystems, Hilborn et al. (6) found only 4 out of 50 cases where growth rates of predators were positively and significantly correlated with forage fish abundance, and 3 cases where there was a significant negative correlation.

Most ecosystems have much richer species composition in forage species communities than the Baltic, and the relative abundance of species (many of which are unexploited) tend to vary greatly over time (Fig. 1) as does their prominence in the diets of predators (12, 13). While specific forage fish species have clearly declined in various locations around the world, those declines were likely offset by other species. In fact, a global analysis by Christensen et al. (14) estimates that forage fish abundance has increased by more than 130% over the last year 100 y, largely due to the decline of higher trophic species that prey on them. However, some marine predators, including seabirds and marine mammals, may be less able to switch between prey as the forage community changes or may be particularly sensitive to local depletion around breeding sites (6). In such cases, spatial controls on catch of specific forage species may be important.

Fig. 1.

Cluster analysis of key forage species in the central California Current ecosystem through 2019. The horizontal lines indicate clusters of typically co-occurring species. The vertical lines indicate temporal shifts in community structure. The colors indicate relative abundance (red, abundant; blue, rare). Reprinted with permission from ref. 13.

It should be noted that, for well-managed piscivore stocks, the analysis by Soudijn et al. (7) does not suggest that piscivores benefit from forage fish exploitation. Their community dynamics modeling analysis shows maximum yields for cod with fishing mortality at or close to zero for forage fish. Benefits for cod accrue only when they are subject to exploitation rates above those that produce maximum sustainable yield. Nor do the authors explore how exploitation of forage fish impacts risk of collapse of forage fish stocks. Essington et al. (15) find that collapse of forage fish stocks is often preceded by heavy fishing pressure, although Szuwalski and Hilborn (16) find only one case in the stocks investigated by Essington et al. with a significant stock–recruitment relationship that would provide a causal link between high exploitation and collapse.

The analysis of Soudijn et al. in PNAS suggests that curtailing or closing down forage fish fisheries could have unintended negative consequences on some predators.

Productivity of forage fish stocks can vary substantially driven by environmental conditions (17) causing stocks’ size to decline and remain well below previously observed levels even with little or no fishing. While caution is clearly warranted when stocks fall to very low levels, the risks of closing or curtailing fishing unnecessarily should also be considered. There is often considerable processing infrastructure associated with fisheries that may not have alternative uses, and closures can result in loss of infrastructure and markets. In addition, there may be knock-on effects on other fisheries. For example, recreational fisheries off California are dependent on sardines for live bait (18), and the lobster fishery in Maine is highly dependent on large quantities of herring to bait traps and potentially increase lobster production (19). Forage fish species appear to vary greatly in relative abundance, their capacity to withstand fishing, and the role they play in their respective ecosystems. As Soudijn et al. (7) demonstrate, ecosystem effects of fishing these stocks can be surprising. EBFM calls for a tailored approach to managing fisheries (for all species) that takes into accounts the unique ecological and human characteristics of each ecosystem, which are still poorly understood in the best of cases (1–3).