Date of Award

Spring 1998

Document Type


Degree Name

Master of Science (MS)


Biological Sciences



Committee Director

Cynthia M. Jones

Committee Member

Douglas S. Vaughan

Committee Member

Anthony J. Provenzano

Committee Member

Mark J. Butler, IV

Call Number for Print

Special Collections LD4331.B46 D67


Separable virtual population analysis (SVPA) models provide estimates of historical recruitment and fishing mortality from analyzing catch data based on the separability of fishing mortality into age specific-gear selection and yearly mortality. These models assume that the catch is randomly sampled and that sampling error is a random variable with constant variance and mean equal to zero. These models do not account for sampling measurement errors that occur when the catch partitioned by age is not measured on every sampling unit without error, and spatio-temporal errors that occur when the observed catch is not representative of the harvested population throughout time and space. I studied the effect of these sampling errors on estimates of recruitment and fishing mortality from the SVPA of Pope and Shepherd (1982) based on the catch (1986- 1996) in the Atlantic menhaden fishery. Underreporting total catch by 20% underestimated recruitment in the same magnitude, but did not result in observable error in fishing mortality. When undercoverage of spatial areas was modeled by a deterministic simulation, average errors in recruitment and fishing mortality were estimated to be from 6% to 116% over the simulated period. Using bootstrapping, I found that stochastic-sampling errors in mean weight also biased estimates of recruitment and fishing mortality, on average up to 8% for the simulated period. In the presence of measurement errors, SVPA underestimated recruitment, but correctly estimated fishing mortality, whereas in the presence of sampling spatial errors SVPA failed to correctly estimate both recruitment and fishing mortality. Thus, sampling errors are potential sources of substantial errors beyond those associated with ageing fish, estimating proportion of fish by age, and improper parametization of a model.


In Copyright. URI: This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).