Modeling Shallow-Water Piscivore-Prey Dynamics in California's Sacramento-San Joaquin Delta

Piscivore-prey dynamics are a key component of understanding ecosystem function and often a key component of fisheries management because many sport fishes are piscivorous. Much research effort has been put into understanding the changing dynamics of the San Francisco Estuary food web, but piscivore-prey dynamics remain poorly understood.  These linkages are of interest to resource managers because of the potential for top-down suppression of ESA-listed native fishes. We used previously published field, diet composition, and bioenergetics modeling data for juvenile (age-1 through age-3) striped bass (Morone saxatilis) and two nonnative prey fishes, threadfin shad (Dorosoma petenense) and Mississippi silverside (Menidia audens). We used nonnative prey as examples because native fishes were eaten too rarely to construct a defensible modeling framework. Our previous research showed striped bass responded differently to these prey during 2001 and 2003, providing us with an empirical contrast. We asked two questions: 1) what is the estimated per capita consumption of shad and silverside, 2) how do variation in overlap of striped bass and their prey, striped bass age structure, and the functional response of striped bass to each prey affect predictions of prey consumption based on predator abundance? We estimated shad consumption was 1-2 orders of magnitude higher than silverside consumption. Median predicted summer-fall shad consumption was 163-3533 grams per striped bass (variation due to predator age); silverside consumption estimates were only 15.6-40.2 grams per striped bass. Thus, shad production had to be considerably higher to outpace striped bass demand. Previous bioenergetics modeling showed striped bass abundance explains most of the variation in population consumption. During 2001 and 2003, striped bass abundance and prey density were weakly correlated across individual sampling events (r = 0.34 and 0.37 for shad and silverside respectively). The correlations between striped bass abundance and prey consumption were higher (0.79 and 0.52 for shad and silverside respectively) due to a combination of bioenergetic allometry and positive density dependence in the prey response equations. Randomly varying striped bass age structure had little effect on consumption estimates. The striped bass-shad correlation was strong enough to suggest that top-down predation pressure on shad is somewhat predictable based on predator abundance, but striped bass abundance could not reliably predict spatial-temporal variation in silverside consumption. Our findings suggest that getting the functional response correct for rare, ESA-listed fishes is a critical step needed to accurately model their susceptibility to top-down suppression by striped bass.