Retrospective analysis of marine biological data from Port Valdez, Alaska: A case study in long -term monitoring

Efforts to understand anthropogenic effects within the Port Valdez study area provide a simple and adaptive model for developing and refining hypotheses to measure the structure of and detect, change in nearshore and benthic habitats. Drivers of change detected by this study are the 1964 Prince William Sound earthquake and the oil transportation and salmon aquaculture industries within the fjord. The study area is a glacial-outwash fjord characterized by strong seasonal and spatial environmental gradients due to glacial influences including seasonally low salinity, high suspended sediment loads, and subsequent high sedimentation rates. Direct and indirect effects from intolerance to low salinity are important in organizing intertidal communities as is habitat structure. Previously unrecognized subtle effects on subtidal fauna from anthropogenic stressors near the marine oil terminal in Port Valdez are identified. Demonstration of statistical methods (variogram estimation, repeated measures analysis of variance, and geostatistical modeling) for field studies with spatially and temporally correlated data should be useful to others seeking to establish new long-term studies or analyze previously collected, long-term field data. Investigation of the re-adjustment of benthic fauna from a large earthquake and ecosystem-level effects of salmon aquaculture are not readily available and this dissertation provides a reference point for any such future studies. Although re-adjustment from the large earthquake was a key process during the study period, salmon aquaculture appears to have a strong effect on the benthic ecosystem.;The model of detecting change is simple and adaptive and provides inputs for larger models and scientific investigations in marine ecosystems. Broad questions are developed through a long-term study of an ecosystem. The hypotheses formulated are then evaluated and refined through retrospective analysis of long-term data and results can be used to refine larger models. This dissertation contributes rigorous, statistically bounded biological time series to regional monitoring programs by providing small-scale, ecological information necessary for larger models. As a result, the information provided in this dissertation should increase the accuracy of ecological models and aid in the management of marine resources.