Date of Award

1992

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Oceanography and Coastal Sciences

First Advisor

Richard E. Condrey

Abstract

Presented in this dissertation are analytical and quantitative methods developed for an experimental gillnet sampling program for the analysis of the dynamics of fish populations, illustrated here on Louisiana's spotted seatrout population. Chapter 1 presents a new method of estimating gillnet selectivity using a system of simultaneous equations and a non-linear iterative maximum likelihood approach. The model solution is a set of parameter estimates which mathematically describe the response surface of capture probabilities for fish of size-class j in mesh size i. As an extension to the nonlinear maximum likelihood gillnet selectivity model, Chapter 2 presents a second approach using the probability of capture to estimate the relative abundances of the various size-classes of fish in the population and develops a variance estimator for the relative abundances. The selectivity model and variance estimation procedures were applied to the experimental gillnet catches of spotted seatrout from 1988 to 1990. In Chapter 3, several multivariate statistical techniques were applied to the experimental gillnet catches of spotted seatrout to examine the distributional ecology of the population. Principal Components Analysis (PCA) revealed a two factor model which explained 78% of the variation in the covariance matrix of the original data. The Principal Component factor loadings were interpreted as recruit (juvenile) and spawner (Adult) abundance for factor 1 and factor 2, respectively. Linear regression analysis showed recruit and spawner abundance had a high positive correlation to salinity during the spawning season, but slopes between the two life history stages of seatrout varied significantly (p $<$ 0.05). After the spawning season (September-December) recruit abundance showed a high negative correlation to salinity. Implications of the non-uniform spatiotemporal distributions of spotted seatrout abundance to management are discussed. Finally, in Chapter 4, a population level assessment is conducted on the spotted seatrout fishery in Louisiana to estimate fishing mortality at age, population size at age, and spawning potential ratios using a Monte Carlo-based Virtual Population Analysis (VPA). Application of the Monte Carlo based simulation (N = 3000) was used to characterize uncertainty in the VPA model output of spawning potential ratios which arise from uncertainty associated with input parameters of natural and terminal fishing mortality.

Pages

169

Share

COinS