Science Collected Works - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemThe functional relationships between temperate fishes and the associated seagrass landscapes(University of Melbourne, 2003)Seagrass occurs commonly in both the intertidal (Zostera mulleri) and subtidal (Heterozostera tasmanica) regions of Port Phillip Bay, Victoria, and may form the primary structural habitat for benthic fish assemblages. Beds of seagrass do not exist as homogenous meadows, but rather are spatially complex, forming patches at a range of scales interspersed within an unvegetated sand matrix. In addition, these beds are temporally variable and susceptible to seasonal reductions and anthropogenic disturbances. The functional relationship between fish assemblages and seagrass areas is also likely to be spatially complex, and will to some degree reflect the underlying spatial structure of the system. In this thesis I took a multi-scaled observational and experimental approach to quantify the functional relationship between demersal fishes and shallow seagrass areas. I first used a spatially explicit approach to quantify which scales appeared to be important in explaining fish-habitat association. I established 185 geo-referenced sampling stations within a 2 x 1 km spatial array, and seined fish and measured depth and habitat variables at each station. To identify the relative importance of habitat (physical and biotic) given the spatial structure of the system, I sequentially modelled the main and interactive effects of spatial location, physical habitat structure, and biotic habitat structure for each species using a combination of regression, autocovariance modelling, and kriging techniques. Depth, seagrass density and seagrass length varied spatially across the array. Both large-scale spatial gradients (alongshore and offshore trends) and short-order spatial autocorrelation were identified. Fishes were generally associated with habitat variables that were intrinsically spatially structured. In addition, one species of pipefish and juvenile King George whiting displayed additional spatial patterns independent of habitat. This initial study descriptively modelled the relationships between fish and seagrass at scales of 0.3 to 2 km, and also identified that the fish-habitat relationship involved a large-scale (>2 km) component, in addition to finer scale (<0.3 km) autocorrelative components. Understanding how and at what scales organisms respond to their environment can provide substantial insight into how subsequent environment changes may alter the assemblage. To identify how larger scaled spatial and temporal trends and habitat parameters interacted to structure the fish assemblage, I undertook a multi-scaled spatio-temporal survey. In this study I sampled both cryptic and motile fishes and recorded the habitat both independent of the fish and directly beneath each fish (latter from diving samples only) from seine and diving samples at 12 sites situated at 2.5 km intervals alongshore, and 4 distance zones located every 250 m offshore. I sampled this array every season, for two years. Habitat, at large scales (km�s), was spatially structured both alongshore and offshore; there was also a small temporal pattern that operated independently of this spatial structure. The distribution and abundance of 90% of the fish species were explained either directly or indirectly by seagrass. A number of species were directly associated seagrass, for example long-dense seagrass. In contrast, other species were associated with the level of seagrass fragmentation rather than the direct presence of seagrass per se. While, habitat variables explained most of the fish assemblage structure, the presence of seagrass alone was not a sufficient predictor of fish, nor was there a typical �seagrass-fish� response. Instead, fish responded to specific habitat classes (e.g., seagrass, patchy seagrass, mixed habitats, or sand areas), but these habitats were often found in specific alongshore and offshore locations, consequently restricting the distributions of the associated fish species. In addition some fish species were also spatially distributed alongshore and/or offshore independent of habitat (e.g., some pipefishes). Temporal patterns were also important, but operated independently of both habitat and space at this scale (km�s). By studying fish-habitat associations at intermediate and fine-scales, I began to identify the specific habitat parameters that fish were actually associated with. At intermediate-scales, I more clearly identified the degree of habitat patchiness fish species were associated with, and the level of habitat selectivity. While, at a fine-scales I identified that the use of microhabitat parameters (patch-interior, edge, gap) by fish species varied. The information of fish-habitat associations at each scale added novel and valuable information, and when combined provided a much greater understanding of how each fish species and the assemblage as a whole responded to the underlying complexity of the habitat. Fish species utilized seagrass areas during different stages of their life history. To identify the population dynamics and possible ontogenetic shifts in habitat use, I measured the size of fish during the spatio-temporal survey. Species abundances varied dramatically between seasons, but were driven by different demographic characteristics. Some species recruited to seagrass areas, but later emigrated to other habitats. Other species recruited to, bred and then senesced in seagrass, while others species (e.g., pipefishes) spent their entire life cycle in seagrass areas. Ontogenetic shifts in habitat use occurred for several species. Fish species utilized different parameters of the seagrass habitat over different stages of their life. For example newly settled seagrass odacids N. balteatus (<2.5 cm) were found in dense seagrass or algae microhabitats, while large individuals were more frequently clumped within edge and gap microhabitats. Similarly, newly settled King George whiting (<3 cm) were associated with seagrass and gaps, 5-9 cm fish were almost only found in gaps within seagrass beds, while larger individuals (> 9 cm) aggregated together in motile schools and were associated with a range of habitats, although gap use was still high. These shifts in habitat association identified that both seagrass cover and habitat fragmentation were important to many fishes but that this importance varied over different phases of their life history. To determine how fish abundance and species richness change in relation to seagrass cover and the level of fragmentation, I experimentally removed different amounts of seagrass (reduced by 16% and 32%) in different spatial configurations (high and low fragmentation) from within four clusters of five 10 x 10 m permanent quadrats. This factorial design enabled disturbance area to be examined independently and interactively with disturbance patch size. Motile fishes were sampled using visual counts while cryptic fishes were non-destructively sampled using a small underwater seine. All fishes were sampled before and three times post-disturbance. Species responded in characteristic ways to experimental disturbance. Seagrass associates (leatherjackets and pipefishes) were either restricted to or more abundant in seagrass microhabitats and consequently abundance decreased with the loss of seagrass. This was contrasted by increased abundance and diversity of edge and gap associates, such as the odacid N. balteatus and King George whiting S. punctata. Gap size was also important. Several species responded to small gaps sizes (4 m2) and the edges of the larger gaps (16 m2) while subadult King George whiting (> 9 cm fishes) exhibited a threshold response, using only large gaps. In additionally, these responses changed with ontogenetic changes. Newly settled N balteatus were found predominantly within the seagrass matrix while larger individuals were found in all microhabitats, commonly focused within and around the edges of gaps, and in contrast to newly settled N. balteatus the abundance of larger individuals increased linearly as seagrass was removed. Despite some assertions, seagrass beds are not homogenous entities. Instead, they form a mosaic that is structured at many different scales. This has important implications for fish communities. In this study I identified changes in the strength and form of fish-habitat associations at scales ranging from meters to kilometers. The observed distribution of fishes in space is a complex array of different habitat elements. There is a deterministic element in which fishes respond to particular biotic habitat components such as seagrass cover or bed patchiness. However, these associations may be modified by the physical structure and the location along the shore of that biotic habitat. Additionally, there is a large-scale temporal dynamic that operates across all scales examined. In other words, the association of a fish with its �preferred� habitat is conditional on a suite of different landscape elements, which may operate additively, or synergistically. This study highlights that measuring the association between organisms and their habitat requires many levels of information, ranging from an understanding of the individual habitat preferences at the fine-scale, to understanding the spatially- explicit structure of fish and habitat at landscape levels. Understanding and predicting fish assemblage structure in the face of habitat change is no simple task, and relies heavily on the integration of fine scale empirical and landscape level studies, but this study demonstrates it is achievable.