Fish Responses to the Introduction of Structural Woody Habitat in Two Coastal Rivers in New South Wales, Australia

Abstract

The declining condition of river systems associated with rapid development of human societies has lead to substantial reductions in the distributions and populations of many freshwater fish. One consequence of the general decline in river condition has been the loss of in-stream Structural Woody Habitat (SWH). Structural woody habitat is now widely recognised as an important component of stream ecosystems, particularly as fish habitat. As a result, there has been an increase in the number of rehabilitation programs that introduce SWH into rivers globally, and in Australia. However, most rehabilitation works remain unmonitored or inadequately monitored and hence their effectiveness remains largely untested and/or poorly recorded. This thesis assesses fish responses to the introduction of SWH in two coastal rivers of New South Wales, Australia. It is focused on evaluation of the key aspects of experimental design that need to be taken into consideration when planning a monitoring program. The results of this study contribute to a growing literature on the monitoring of fish responses to rehabilitation works using introduced SWH. In 2000 a reach scale experiment was undertaken in the Williams River, NSW, with the introduction of SWH to a test reach compared with a geomorphologically similar control reach. Changes in fish species richness, abundance and assemblage structure were quantified over a period of five years after treatment using a Before-After-Control-Impact-Paired (BACIP) design and analysis. The initial increase in fish species richness observed in the two years following SWH introduction (reported in a previous assessment) appeared to have been sustained three years on (i.e. five years after SWH introduction). This outcome suggests that the increased habitat diversity in the test reach following SWH introduction was sufficient to elicit a detectable and sustained response by fishes over the study period of five years. However, the increased fish abundance noted two years after SWH introduction into the test reach appears to have dissipated over the following three years. Analysis of abundance, excluding a small highly abundant species, Retropinna semoni (Australian smelt), suggested that the initial increase in the test reach was due solely to this species. The most plausible explanation for the decrease in fish abundance after the first two years of the experiment was the decrease in available SWH over the study period due to burial by substrate mobilized during high flow events. The Williams River study is representative of the few rehabilitation projects that attempt to monitor fish responses to SWH introduction and, as such, provided the opportunity to examine inadequacies in experimental design and data analysis. Specifically, the Williams River experimental design suffered from: unbalanced sampling effort between the test and control reaches, irregular sampling, a lack of spatial and temporal replication, and potential spatial autocorrelation. The early success reported for the reach scale experimental introduction of SWH in the Williams River motivated the scaling up of rehabilitation works into a larger river, the Hunter River. Aspects of the Hunter River study that improved upon the Williams River study were: increased replication of treatments and controls that were similar at the start of the experiment, increased temporal replication, balanced sampling effort across all sites, systematic temporal sampling, and an effort to measure the potential for spatial autocorrelation. Fish population responses at the meso-habitat scale (i.e. in riffles and pools) were tested using a Multiple-Before-After-Control-Impact (MBACI) experimental and analytical design. The introduction of SWH into riffles and pools in the Hunter River did not elicit a strong response from fish populations at the meso-habitat scale. In the riffle experiment, the introduction of SWH (deflector jams) appeared to create areas of slack water habitat which were utilised by one native fish species (R. semoni) and one exotic species (Gambusia holbrooki - mosquito fish). No increase in fish species richness or abundance following SWH introduction in treatment pools was detected by the MBACI analysis. The most likely explanation for the lack of response to SWH in Hunter River pools is that the volume of wood introduced was very low (approximately 90% less) relative to natural wood levels in this and other Australian rivers, and therefore insufficient to elicit a measurable response from the fish assemblage at the pool scale...