Using Age-Specific Abundance Indices to Disaggregate the Forces Driving Longfin Smelt Production in California's San Francisco Estuary

In California’s San Francisco Estuary, longfin smelt Spirinchus thaleichthys, show a persistent association between freshwater flow and Age-0 abundance.  To clarify this and other aspects of longfin smelt population dynamics, we developed a simulation-based approach based on ten alternative spawner-recruit models derived from four easily defensible assumptions: (1) a Ricker spawner-recruit model provides a sufficiently flexible framework; (2) longfin smelt live two years; (3) freshwater flow affects recruitment; (4) production per unit of flow has declined.  We applied these assumptions to four pairs of conceptual alternatives: (1) recruits per spawner is a log-linear function of freshwater flow versus recruits per spawner is a nonlinear function of  freshwater flow; (2) ‘spawner’ abundance can be represented using Age-0 fish indices versus spawner abundance should be represented explicitly using indices of Age-2 fish; (3) Age-0 production is density-independent versus density-dependent; and  (4) spawner-recruit parameters did not change after the 1986 invasion of the estuary by overbite clam Potamocorbula amurensis, versus they did.  We found that accounting for survival between the juvenile and adult life stages was the most important factor affecting the accuracy of the simulated time series of longfin smelt relative abundance followed distantly by how we modeled recruits per spawner.