Pacific Salmon Environmental and Life History Models: Advancing Science for Sustainable Salmon in the Future
Importance of Marine-Derived Nutrients in Establishing Escapement Goals for Pacific Salmon
David R. Bernard and Robert A. Clark
Abstract.—Traditional methods were used to estimate optimal escapement goals (goals that produce maximum sustained yield) for Pacific salmon Oncorhynchus spp. in a modeling experiment that included postulated depensatory effects of marine-derived nutrients (MDNs) on production. Depensatory and compensatory density dependence in survival rates of young salmon were modeled as a function of parental abundance (the MDN model). Parameters in the MDN model were changed to produce study cases with different strengths of depensation at low numbers of parents, different trajectories in the depensatory function, and different base productivity in salmon populations. Stronger depensation tended to produce an Allee effect (production below replacement at low abundance) as an emergent property while depensatory functions with flatter (lower) trajectories depressed carrying capacity. Several study cases involving stronger depensation, flatter trajectories, and/or lower base productivity had unrealistic production curves (production below replacement). Optimal escapement goals and optimal harvest rates were estimated for study cases having realistic production curves using three traditional methods: halving an estimate of carrying capacity (used when collected data came from a period of negligible exploitation rates), regressions (for data taken when exploitation rates have been moderate), and with a function of estimated intrinsic rates of increase (when exploitation rates have been high). Comparisons of estimates against known values from a deterministic version of the MDN model indicated that, in all study cases except one, goals estimated with traditional methods produced expectations of nearly optimal fishing to recruitment underfishing. The one exception concerned halving the estimated carrying capacity for a population for which depensation from MDNs was strong with a flat trajectory. In that case, simulations with a stochastic, dynamic version of the MDN model that included process error showed this population would likely not exist. Additional stochastic simulations also reinforced conclusions from the deterministic model, except that regression often failed because of lack of fit when depensation from MDNs was moderate or strong for populations with low base productivity. Hydrological and biological characteristics that relate to modeled effects of MDNs on salmon production were postulated.