Some Effects of Parameter Uncertainty in Density-Independent and Density-Dependent Leslie Models for Fish Populations

Abstract

Density-independent and density-dependent Leslie models were investigated by Monte Carlo methods. Random values for parameters of striped bass (Morone saxatilis), white perch (Morone americanus), and tomcod (Microgadus tomcod) populations were selected from truncated normal distributions with standard deviations equal to 10% of the mean. Only total population size after 40 yr was considered. The error propagation properties of the density-independent models are strongly influenced by model assumptions (e.g. calculating egg to 1-yr-old survival to ensure an eigenvalue of 1.0) and by the way model parameters are estimated (e.g. reestimated from data each year). Prediction errors on total population size depend on the number of age-classes in a species, but become insensitive when the number of classes exceed 7. Under the very restrictive assumptions used here, there is little difference in the error propagating properties of alternative density-dependent models.Key words: matrix, population, striped bass, white perch, tomcod