Date of Award


Degree Type


Degree Name

Master of Science


Faculty of Science, Technology and Engineering

First Advisor

Charles Jacoby

Second Advisor

Pierre Horwitz


As we attempt to maintain marine biodiversity mainly by focussing on habitats, we need to understand how marine biodiversity is affected by seagrass loss. Although managers and researchers widely acknowledge that habitat loss results in changes to marine biodiversity, quantitative knowledge of these changes is generally poor. In this study, fish assemblages (as one component of the biodiversity of sandy beaches) were examined in Cockburn Sound, Western Australia, to assess patterns that may be related to presence or absence of adjacent seagrass beds, If consistent patterns are evident, they may enable predictions regarding the effects of seagrass loss on the fish assemblages. Prior to examination of this main question, seagrass and bare sand habitats were sampled as part of a pilot study. Analyses of the data collected concentrated on examining the biases and variability associated with different sampling gear, and changes in the precision of estimates derived from different levels of replication. Further consideration was given to reducing the relative weighting of very numerous species in analyses by examining the effects of data transformation. Results suggest that researchers seeking to detect a 'signal' of environmental change amid the 'noise' that results from variability in catches and the numerical dominance of a few species should select methods of sampling, levels of replication and types of data transformation with an understanding of the associated influences. The main sampling program was conducted in May-June and October November 1995 (months identified as suitable pre- and post-recruitment periods), over six beaches in Cockburn Sound. Assemblage and population level analyses indicated that consistent trends related to the presence of adjacent seagrass were not evident, as differences among beaches was the dominant spatial trend. Several environmental variables appeared to influence these differences. The most significant were water depth, wind, amount of drift seagrass and wave exposure. Degree of exposure and amount of drift seagrass probably had the greatest structuring effects on the species assemblages. Both were determined by the position of habitat patches in relation to wind direction and proximity of other habitats, Although patterns associated with the presence or absence of adjacent seagrass Were not clear, these findings indicated that seagrass beds had considerable influence on the fish assemblages of some beaches, Populations of some species changed between May-June and October November due to recruitment of juveniles to the beaches. Although juveniles of some species use other habitats, there is a strong suggestion that the beaches of Cockburn Sound are regionally important in the ecology of species such as Aldrichetta forsteri, Ammotretis elongatus, Sillago uittata and Sill ago schomburgkii. Seagrass loss in Cockburn Sound is unlikely to have affected the more mobile fish species, but is likely to have caused a decline in the populations of less mobile, site-associated species of both seagrass and sand habitats. Additional loss of seagrass within Cockburn Sound is likely to further reduce the populations of many species. If these components of biodiversity are to be maintained, then management of marine biodiversity in the region needs to be conducted with a recognition of the differences in assemblage composition among habitat patches. Conservation of just a few habitat patches will not adequately represent the full range of species and age classes present.