For my Sea Grant/NMFS population dynamics graduate fellowship I spent three summers as an intern with Dr. Alec MacCall at the Southwest Fisheries Science Center Santa Cruz Lab. During that time we investigated the problem of model uncertainty in the stock assessment. I was particularly concerned the population biology of the slow growing genus Sebastes.
To understand the influence of model uncertainty, I developed a population simulation model. I fit population models to the simulation model and compare parameter estimates to the "true" simulation values [3]. I also compare goodness of fit measures. The simulation model was parameterized to mimic the biology of widow rockfi�sh (Sebastes entomelas).
There are three sources of stochasticity in the simulation model. The model has stochastic recruitment, stochastic natural mortality and stochastic individual growth. Stock assessments routinely estimate stochasticity in the recruitment function [5]. However, it is very rare to estimate uncertainty in the natural mortality rate. And, to the best of my knowledge, no stock assessment has assumed stochasticity in individual growth.