James Tweedley

The work for this thesis was undertaken in Broke Inlet, a seasonally-open estuary on the south coast of Western Australia and the only estuary in that region which is regarded as “near-pristine” (Commonwealth of Australia, 2002). The only previous seasonal studies of the environmental and biotic characteristics of this estuary involved broad-based descriptions of the trends in salinity, temperature and ichthyofaunal characteristics at a limited number of sites. Furthermore, no attempt has been made to identify statistically the range of habitats present in the nearshore and offshore waters of this system, and the extents to which the characteristics of the fish and benthic invertebrate faunas are related to habitat type. These types of data provide not only reliable inventories of the habitat and faunal characteristics of Broke Inlet, but also a potential basis for predicting the likely impact of anthropogenic and climatic changes in Broke Inlet in the future. The main aims of this thesis were as follows. (1) To use the method of Valesini et al. (2010), which employs enduring environmental characteristics, to identify quantitatively the range of habitats present throughout the nearshore and offshore waters of Broke Inlet. The enduring environmental characteristics represent three broad categories, i.e. the location of any site in terms of its proximity to marine and freshwater sources, the degree of exposure to wave activity and the type of substrate and/or submerged vegetation. (2) To test the hypothesis that the species richness, density, diversity and species compositions of the fish and benthic macroinvertebrate faunas differ among habitat types, seasons and, in the case of the fish fauna, also years. (3) To test the hypothesis that the pattern of relative differences among habitat types, as exhibited by their faunal compositions, is correlated with that defined by their (i) enduring environmental characteristics and (ii) non-enduring environmental characteristics (water physico-chemical variables and sediment characteristics). A high resolution satellite image and a digital elevation model of Broke Inlet were used to measure the enduring environmental characteristics at 104 and 36 widely-distributed sites in nearshore and offshore waters, respectively. These data were used to construct separate Manhattan distance matrices for nearshore and offshore waters, which were then subjected to the CLUSTER and SIMPROF routines in PRIMER v6 to identify the various groups of sites that did not differ significantly in their environmental characteristics and which were thus considered to represent habitat types. Twelve and four distinct habitat types were identified in nearshore and offshore waters, respectively. The ichthyofaunas at sites representing 11 nearshore (A-K) and three offshore (A-C) habitat types were sampled seasonally for two consecutive years using seine and gill nets, respectively. A total of 83,047 fish was collected from nearshore waters, representing 27 species from 19 families, with 99.6% of those fish belonging to six species which represent the Atherinidae (Atherinosoma elongata, Leptatherina wallacei and Leptatherina presbyteroides) or Gobiidae (Afurcagobius suppositus, Pseudogobius olorum and Favonigobius lateralis) and complete their life cycles within the estuary. Each of these species were found at each nearshore habitat type, except for P. olorum, which was not caught at habitat A. The species richness, density and diversity of the nearshore fish fauna differed significantly among habitats, seasons and years, with habitat being the most influential factor. Generally, mean species richness and density were greatest at habitat types located in the entrance channel (B, G and H) and/or on the southern shore of the basin (C and G), while the fish assemblages were most depauperate at habitats near freshwater sources (A and J). The nearshore ichthyofaunal composition of Broke Inlet differed significantly among habitats, seasons and years, with the first again being the most influential. However, the extents of the overall differences in composition during each sampling occasion were moderate. The lack of very pronounced ichthyofaunal differences among the various habitat types reflects the widespread distributions and high abundances of the above atherinid and gobiid species, and particularly of A. elongata and L. wallacei, which typified the fish fauna of each habitat type on almost all sampling occasions. The most distinctive of the faunas were those at habitat types in the entrance channel, where L. presbyteroides and F. lateralis, which are typically found in higher salinities, were in their greatest densities, and where several marine straggler species, such as the labrids Notolabrus parilus and Achoerodus gouldii, were occasionally caught. Seasonal and inter-annual changes in ichthyofaunal composition were small and often not consistent across habitats. The pattern of relative differences among nearshore habitats in terms of their ichthyofaunal composition was correlated at a moderate level with that defined by their enduring environmental characteristics in all but one of the eight seasons. Such indicate that the ichthyofaunal composition likely to be found at any site in the nearshore waters of Broke Inlet at any time of year can be predicted, simply by assigning that site to its most appropriate habitat type on the basis of its enduring environmental characteristics. Differences in ichthyofaunal composition among habitat types were also correlated, but to a slightly greater extent, with those among the suite of non-enduring water physico-chemical variables, with salinity and the biomass of macrophytes being particularly relevant. Gill netting in the three offshore habitats yielded 1,050 fish representing 31 species. Species richness, catch rates and diversity all varied significantly among habitats, with the values for each of these biotic characteristics always being greatest at habitat A in the entrance channel and lowest at habitat B near the Shannon River mouth. These biotic variables did not always vary, however, among seasons and/or years. In contrast to the situation in nearshore waters, the offshore ichthyofauna comprised mainly marine estuarine-opportunists and marine stragglers, which contributed 84% to the number of species and 80% to the total number of fishes. The contribution of individuals belonging to the marine straggler guild was only 5% and no estuarine resident species were caught. Ichthyofaunal composition in offshore waters differed significantly among habitats, seasons and years, with habitat being the most influential factor. Faunal composition only differed among habitats in spring and autumn, and even then the extent of those differences was low. During those seasons, habitat B contained the most distinct and depauperate fauna, which was typified mainly by Mugil cephalus and Aldrichetta forsteri. In contrast, the fish assemblages at habitats A and C were also typified by Arripis georgianus, Arripis truttaceus, Rhabdosargus sarba, Pagrus auratus, Pseudocaranx dentex and Engraulis australis. The pattern of relative differences among offshore sites in terms of their ichthyofaunal composition was significantly correlated with that defined by their enduring environmental characteristics only in autumn, but was moderately correlated with that exhibited by the suite of non-enduring water physico-chemical variables in each season except summer. Seasonal sampling of benthic macroinvertebrates at six of the nearshore habitat types (A, C, D, F, H and K) for a year yielded 7,485 individuals representing 28 species and seven phyla and, at the three offshore habitat types (A-C), 2,459 individuals representing 26 species and eight phyla. Polychaetes (64 and 57%) and crustaceans (24 and 34%) were the most abundant taxa in nearshore and offshore waters, respectively. The mean density of invertebrates in the nearshore waters did not differ significantly among habitats, but did vary significantly among seasons, and was greatest in summer. The mean densities of invertebrates in offshore waters did not differ significantly, however, among habitats or seasons. The compositions of nearshore benthic macroinvertebrate assemblages differed significantly among habitats and, less conspicuously, seasons. Comparisons between the faunal compositions in each pair of habitats in spring and summer were almost invariably significantly different and to a moderate extent. However, such pairwise comparisons were rarely significant in autumn and winter. Habitats A and K contained the most distinct and depauperate invertebrate fauna, comprising mainly the polychaete Capitella capitata and amphipod Corophium minor, whereas the other habitats also contained large numbers of the polychaete Ceratonereis aequisetis. In offshore waters, the composition of the benthic macroinvertebrate assemblages differed to a low to moderate degree among habitats, with habitat B containing the most distinct fauna due to large densities of C. minor. The extent of seasonal differences in these faunal compositions was small. The pattern of relative differences among nearshore habitats in terms of their benthic macroinvertebrate composition was highly correlated with that defined by both their (i) enduring environmental characteristics and (ii) non-enduring water physico-chemical and sediment characteristics in spring and/or summer. Although the faunal and enduring environmental data were not correlated at offshore sites, the fauna and non-enduring environmental variables at those sites were correlated to a moderate extent. In summary, the composition of the nearshore fish fauna at any site in Broke Inlet at any time of year can now be predicted by allocating that site to a particular habitat type on the basis of its enduring environmental characteristics. The less consistent spatial correlations between the compositions of the offshore fish fauna and benthic macroinvertebrate faunas and the enduring environmental variables largely reflected the ubiquitous nature of the majority of the abundant species representing those faunas, i.e. they are typically at least moderately abundant in all habitats and thus have no strong preferences for a particular habitat type.