Date of Award

Spring 1994

Document Type


Degree Name

Doctor of Philosophy (PhD)


Ocean/Earth/Atmos Sciences



Committee Director

Phillip R. Mundy

Committee Member

Edward Markowski

Committee Member

John R. McConaugha

Committee Member

Chester E. Grosch


Harvest control in salmonid fisheries was examined as a problem in the formulation of regulations which restrict time and area of fishing. An ability to rigorously define and compare the form of the progression of migration across time and between harvest areas was judged fundamental to objective harvest decisions. Identification and evaluation of statistical methods appropriate to the comparison of empirical migratory time densities was performed.

The development of the measure of central tendency (mean date) of the time density as the consistent, unbiased estimator of migratory behavior was given. Practical evidence demonstrated that the mean date was highly resistant to factors which contribute variability to the basic expression of migratory behavior.

Brood year cycle and locality were treated as fixed effects in statistical analyses applied to the timing statistics of catch and spawning escapement. Fixed effects analysis of variance models were examined to analyze differences in mean dates of migration. Multiple comparison analysis, Scheffe's a priori method, correlation, and multiple regression analyses were employed to objectively define the performance of the fishery and the escapement in time and space.

Highly significant differences were shown to exist between timings of odd and even populations. For each cycle year for both catch and escapement the management districts were shown to be highly distinct with respect to timing behavior. Linear combinations of districts also demonstrated distinctive timing behavior. In all cases, the subset of districts explained a large percentage of total sound-wide variation in the mean dates of migration for both cycle years and data categories.

Migratory timing as a quantitative description of migratory behavior was rigorously compared across years and harvest areas. Results of the analysis of even and odd cycles were consistent with the genetic distinctness between these two populations, and with the hypothesis of the genetic heritability of migratory timing.

Several statistical models were shown to be extremely robust for determining differences in migratory behavior when the measures of central tendency of the time densities were employed as modeled variables. It was concluded that this system of analysis will serve as a design standard for the comparison of migratory behavior and will be applicable to the needs of harvest control for any migratory organism.