The uncertainty in estimates of the production of larval fish derived from samples of larval abundance

Publisher Website
Google Scholar
Abstract
A Monte Carlo simulation model was used to investigate uncertainty in estimates of the production of newly hatched larvae by a fish population when the estimates are derived from samples of larval abundance. The model simulates (1) hatching, growth, and mortality of a population of larvae; (2) a series of cruises which sample the population; (3) back-calculation to estimate the number hatched; and (4) comparison of estimates of the production of hatched larvae (LP). The model is applied for several combinations of (a) the time interval between cruises; (b) the coefficient of variation of larval concentration within a cruise (CVWC), which depends on sample size; and (c) for several other variables which characterize the simulated population. The coefficient of variation of estimates of the number of larvae produced (CVLP) increases sharply if the interval between cruises is greater than twice the standard deviation of the temporal distribution of the larval hatching. It is equally sensitive to CVWC. For example, doubling sample size by increasing the number of samples within cruises or by halving the interval between cruises has a similar effect. Of the random variables that characterize the population, CVLP is most sensitive to variability in growth rate. There is a bias in the estimate of larval production if mortality rate is length-dependent, or if estimates of growth rate used in calculations are incorrect. The simulations show that estimates of larval production are inherently uncertain.

This publication has 0 references indexed in Scilit: