Analysing the effects of anthropogenic activities on two aquatic ecosystems in Western Australia and identifying sustainable policies for ecosystem-based management
Fretzer, Sarah (2013) Analysing the effects of anthropogenic activities on two aquatic ecosystems in Western Australia and identifying sustainable policies for ecosystem-based management. PhD thesis, Murdoch University.
Anthropogenic impacts such as fishing and eutrophication are significant challenges to the sustainable management of aquatic ecosystems. This study used two ecosystem modelling techniques to investigate the effects of fishing and eutrophication on aquatic ecosystems in Western Australia.
Firstly, a qualitative modelling technique called ‘loop analysis’ or ‘qualitative modelling’ was used to characterise the dynamics of the seagrass ecosystem in Shark Bay, Western Australia (Chapter 2). A qualitative model based on differential equations, was developed to represent the dynamics of the seagrass ecosystem, particularly interactions among tiger sharks, megafauna (e.g. dugongs), and megafauna prey (Fig. 2.2). Although the model structure generated some uncertainty about model predictions and model stability, it was possible to assess the stability of the model and to determine the response signs of model variables by applying data and magnifying loops.
Qualitative modelling analyses indicated a strong top-down control by tiger sharks and suggested that this controlling effect occurred in four stages. A step-by-step increase in tiger sharks (States 1 and 2) led to a habitat shift by the megafauna out of seagrass meadows and into safer, deeper channel habitat. A step-by-step decrease in tiger shark numbers led to the megafauna returning to seagrass meadows, leading to a decrease of megafauna prey in this habitat (steps 3 and 4).
Thus, tiger sharks influenced the use of seagrass habitats by megafauna species through direct and behaviourally mediated impacts. Further, megafauna responses to tiger shark predation risk established alternating predation pressure on different prey groups within seagrass habitats. Curiously, despite the fact that only some megafauna species (e.g. dugongs) are major components of diet of the adult tiger sharks, the perceived predation risk created by the high abundance of of tiger sharks in summer appears sufficient to cause megafauna species to leave (or under-utilise) feeding habitats in seagrass meadows. Thus, the modelling results suggest that the abundance of tiger sharks exerts an important top-down, regulatory influence on the other ecosystem components.
This regulatory system has the potential to become imbalanced if there is a decrease in the abundance of adult tiger sharks in Shark Bay, as has occurred tiger shark populations in other areas worldwide. Targeting of tiger sharks by fishers in the waters of Northern Australia and Indonesia has increased steadily during the last years and impact the tiger shark stock in Shark Bay if, as has been hypothesized, there is single common stock. A qualitative trophic model suggested that the activities of recreational fishermen within Shark Bay reduce prey availability for juvenile tiger sharks, an impact which might adversely affect the tiger shark population and, thus, the dynamics of this seagrass ecosystem.
Evidence of the ecological importance of tiger sharks and the potential impact of a population decline emphasises the need to sustain the tiger shark population in Shark Bay.
In the second part of this study, a quantitative modelling technique, Ecopath with Ecosim and Ecospace, was applied to the ecosystem of the Peel-Harvey Estuary, Western Australia. A key impact on this ecosystem is the the Dawesville Channel, an artificial entrance channel was constructed in the mid-1990s to increase the flushing and reduce nutrient concentrations in the estuary.
Ecopath was used to analyse the impact of the Dawesville Channel on the estuarine ecosystem. A large dataset was collected for model development, a process that uncovered significant data gaps (e.g. missing data on detritus pool and dietary information and indicated important areas for further research (Chapter 3).
Two identical Ecopath models (comprising 30 living functional groups) were otherwise developed for the Peel-Harvey Estuary to describe the state of the ecosystem before (‘pre DC’) and after (‘post DC’) the opening of the Dawesville Channel (Chapter 4). Modelling found that, in addition to changes in the community structure of plants, fish and invertebrates, the entire ecosystem of the Peel-Harvey Estuary has declined drastically in total biomass since the opening of the Dawesville Channel, as has the biomass at each trophic level and in the size of flows between the functional groups. Changes in flows and transfer efficiencies suggested a change in the functioning of the ecosystem in which consumption has become a more important and more efficient flow since the opening of the channel. Analysis of network and system statistics indicated that food web structure had also changed, with more linkages in the ‘post DC’ model and thus a more web-like structure than in the ‘pre DC’ model. Modelling also identified changes in cycling processes and suggested that the ecosystem in the ‘post DC’ model was not able to keep carbon within the system, even though: (i) the food web has developed more linkages and (ii) with less primary production and less cycling, the size of the ecosystem has decreased drastically since the channel opening.
Overall, the results of the Ecopath modelling indicated that the Dawesville Channel has markedly impacted the features, functioning and services of the Peel-Harvey Estuary (Chapter 4). Several indices were applied that suggested that both the ‘pre DC’ and the ‘post DC’ models were highly immature. Ecopath was also applied to investigate the impact of the Dawesville Channel on ecosystem services. Ecopath modelling indicated that all ecosystem services had declined, such as provisioning services (catches), regulating services (CO2-Fixation) and supporting services (nutrient cycling, primary production and biodiversity). Unfortunately, it was not possible to locate data relating to cultural services (tourism) for the ‘pre DC’ model.
To support the reliability of the Ecopath and Ecosim predictions, model uncertainty and the sensitivity of the parameter settings were assessed in detail (Chapter 5). Overall, the results of this analysis indicated that the parameter settings for the ‘pre DC and ‘post DC’ models were robust and did not lead to uncertainties regarding modelling results and predictions. However, the vulnerability settings are crucial for Ecosim and Ecospace and need to be treated with caution.
Ecosim was applied to identify: 1) the impact and effectiveness of the selective reduction of different primary producers and 2) the impacts of fishing on target and non-target species on the ecosystem model (Chapter 6). The application of Ecosim requires fitting a model to time series data; for this study, the sourcing and fitting of time-series data indicated the importance and uncertainty of vulnerability settings. Three categories of vulnerabilities were identified: (a) vulnerabilities that did not have any effect on time series fitting (category 1); (b) interactions in which the lowest sums of squares occurred at low vulnerability settings (v=1 or 2, category 2); and settings that had a drastic impact on model fitting (category 3).
The Ecosim simulations indicated that fishing affected almost all functional groups in the model, not just the target species. The recreational fishing sector also had a very strong impact on many functional groups, particularly Blue Swimmer Crabs and other invertebrate groups like bivalves and gastropods. The commercial fishing sector affected functional groups less than the recreational sector, but affected a range of estuarine fish groups including non-target fish species. Thus, the results of this study suggest that it may be not advisable to close those fleets completely as some aspects of the estuary ecosystem appear to benefit from increasing fishing pressure. Some fish groups and some target species responded positively to the closure of certain fleets, while others – particularly waterbirds and other top predators – did not (Table 6.8). Ecosim analyses highlighted the need for more data to ensure sustainable management, but suggested that the coexistence of fleets might be a better solution for sustaining catches and group biomasses in the future.
Ecosim modelling indicated that selective plant removal is a reasonable management tool for this estuary. However, nutrient reduction and, thus, the permanent reduction of microscopic algae appears to be more ecologically and economically worthwhile (Fig. 6.12). Removing aquatic plant groups showed no significant longterm change in biomasses and the magnitude of short-term effects was much higher than for long-term effects. The Ecosim simulations demonstrated that only a permanent reduction in microscopic algae led to a reduction in total biomass. Reducing phytoplankton might be worthwhile because, although blooms of Nodularia spumigena no longer occur in the estuary because the salinities are too high (Huber, 1985), the estuary now contains several phytoplankton species (e.g. Heterosigma akashiwo) that cause blooms in other ecosystems (Guiry & Guiry, 2010). While the effects of phytoplankton blooms on the ecosystem depend on the size (i.e. the biomass) of the bloom, even blooms that only double the biomass of microscopic algae can have drastic long-term effects. This study supports the conclusion that a reduction in phytoplankton through management of nutrient input in the estuarine catchment represents the only ecological and economical management scenario that provides long-term sustainability for this ecosystem (Chapter 6).
Ecospace modelling represents biomass dynamics over two-dimensional space and time. For this study, a model with fours habitats (shallow mud, deep sand, rocks and plant habitat) was developed. By applying Ecospace, the effects of reducing plant habitat and the effectiveness of two Marine Protected Areas were investigated, with specific consideration of waterbirds (Chapter 7). The Ecospace simulations suggested that waterbirds and piscivorous waterbirds were impacted by fishing and would benefit slightly from an introduction of a MPA, in particular a MPA at Point Grey. Further, the results of this Ecospace scenario indicated that waterbirds would profit from the reduction of plant habitat, whereas piscivorous waterbirds showed a small decline in biomass after removal of aquatic plants. Under the current fishing effort, the total biomass of the system and of the fish community increased. Thus, while the major prey groups of piscivorous waterbirds increased in biomass, but piscivorous waterbirds did not benefit from increased prey biomasses in the model, presumably because of the competition for fish. Ecospace modelling indicated that the catches would also increase drastically and, thus, that piscivorous waterbirds were in direct competition with the fishing sectors and other piscivorous predators (e.g. dolphins and sharks) and were out-competed for fish. The modelling suggest that the sustainable management of the fishing sectors is essential for bird conservation.
A MPA at Peel Inlet led to lower catches under the current fishing scenario and catches declined even further under lower fishing effort. In contrast, after introducing a MPA at Point Grey, the total catch only declined when the fishing effort was lessened. The Ecospace simulations indicated that an MPA at Point Grey increased the biomasses of functional groups and target species and also raised the total biomass of the system; however, these effects strongly depend on fisheries management (Chapter 7).
Overall, the qualitative and quantitative modelling methods applied in this study improved our understanding of the dynamics and functioning of the Shark Bay and Peel-Harvey ecosystems (Chapter 8). Both approaches produced robust and reliable results. If precise quantitative predictions are required for a management scenario, Ecopath with Ecosim is the appropriate method to choose, as this approach can deliver detailed changes in biomass and catches. In contrast, qualitative modelling only indicates the direction of change, which might not always satisfy management needs. However, qualitative models are the ideal method when management decisions have to be made fast and when a detailed data set of the ecosystem is not available.
|Publication Type:||Thesis (PhD)|
|Murdoch Affiliation:||School of Veterinary and Life Sciences|
|Item Control Page|
Downloads per month over past year