| Abstract/Description or Keywords |
It is often recognized that a quantitative assessment of the cumulative effects, both additive and non-additive, of multiple stressors would provide a more realistic representation of the factors that influence sockeye salmon (orhynchus nerka) migration mortality. Towards achieving this type of assessment, this research project first presents a literature review of multivariable methods currently applied in Fraser River sockeye salmon migration studies and in the fields of fisheries, biology, and medicine which could be used to analyze cumulative effects. Papers taken solely from Fraser River sockeye salmon research revealed a limited number of multivariable methods being applied and the sub-optimal reliance on univariable methods for multivariable problems. The review of fisheries and biological science literature identified a number of additional methods for dealing with cumulative effects while the review of medical science literature did not reveal any additional methods. The literature review also presents a guide for how to apply each of these methods to other cumulative effects studies and more specifically how to apply them to study Fraser River sockeye salmon migration survival. The second part of this project presents an application of two of these multivariable methods, regression trees and random forests, to describe and predict the cumulative effects of multiple habitat and stressor variables on Fraser River sockeye salmon prespawn mortality (PSM). The results of this analysis show that although a number of these variables may relate to sockeye salmon PSM, only a few variables representing the timing of entry into the Fraser River, the destination spawning ground, and human population density are required to predict Fraser River sockeye PSM. |
