Resolving ecosystem complexity in ecological risk assessment for mine site rehabilitation.

CONTEXT: Ecological Risk Assessments (ERAs) are important tools for supporting evidence-based decision making. However, most ERA frameworks rarely consider complex ecological feedbacks, which limit their capacity to evaluate risks at community and ecosystem levels of organisation. METHOD: We used qualitative mathematical modelling to add additional perspectives to previously conducted ERAs for the rehabilitation of the Ranger uranium mine (Northern Territory, Australia) and support an assessment of the cumulative risks from the mine site. Using expert elicitation workshops, separate qualitative models and scenarios were developed for aquatic and terrestrial systems. The models developed in the workshops were used to construct Bayes Nets that predicted whole-of-ecosystem outcomes after components were perturbed. RESULTS: The terrestrial model considered the effect of fire and weeds on established native vegetation that will be important for the successful rehabilitation of Ranger. It predicted that a combined intervention that suppresses both weeds and fire intensity gave similar response predictions as for weed control alone, except for lower levels of certainty to tall grasses and fire intensity in models with immature trees or tall grasses. However, this had ambiguous predictions for short grasses and forbs, and tall grasses in models representing mature vegetation. The aquatic model considered the effects of magnesium (Mg), a key solute in current and predicted mine runoff and groundwater egress, which is known to adversely affect many aquatic species. The aquatic models provided support that attached algae and phytoplankton assemblages are the key trophic base for food webs. It predicted that shifts in phytoplankton abundance arising from increase in Mg to receiving waters, may result in cascading effects through the food-chain. CONCLUSION: The qualitative modelling approach was flexible and capable of modelling both gradual (i.e. decadal) processes in the mine-site restoration and the comparatively more rapid (seasonal) processes of the aquatic ecosystem. The modelling also provides a useful decision tool for identifying important ecosystem sub-systems for further research efforts.

Optimising Land-Sea Management for Inshore Coral Reefs.

Management authorities seldom have the capacity to comprehensively address the full suite of anthropogenic stressors, particularly in the coastal zone where numerous threats can act simultaneously to impact reefs and other ecosystems. This situation requires tools to prioritise management interventions that result in optimum ecological outcomes under a set of constraints. Here we develop one such tool, introducing a Bayesian Belief Network to model the ecological condition of inshore coral reefs in Moreton Bay (Australia) under a range of management actions. Empirical field data was used to model a suite of possible ecological responses of coral reef assemblages to five key management actions both in the sea (e.g. expansion of reserves, mangrove & seagrass restoration, fishing restrictions) and on land (e.g. lower inputs of sediment and sewage from treatment plants). Models show that expanding marine reserves (a 'marine action') and reducing sediment inputs from the catchments (a 'land action') were the most effective investments to achieve a better status of reefs in the Bay, with both having been included in >58% of scenarios with positive outcomes, and >98% of the most effective (5th percentile) scenarios. Heightened fishing restrictions, restoring habitats, and reducing nutrient discharges from wastewater treatment plants have additional, albeit smaller effects. There was no evidence that combining individual management actions would consistently produce sizeable synergistic until after maximum investment on both marine reserves (i.e. increasing reserve extent from 31 to 62% of reefs) and sediments (i.e. rehabilitating 6350 km of waterways within catchments to reduce sediment loads by 50%) were implemented. The method presented here provides a useful tool to prioritize environmental actions in situations where multiple competing management interventions exist for coral reefs and in other systems subjected to multiple stressor from the land and the sea.

A database of marine phytoplankton abundance, biomass and species composition in Australian waters.

There have been many individual phytoplankton datasets collected across Australia since the mid 1900s, but most are unavailable to the research community. We have searched archives, contacted researchers, and scanned the primary and grey literature to collate 3,621,847 records of marine phytoplankton species from Australian waters from 1844 to the present. Many of these are small datasets collected for local questions, but combined they provide over 170 years of data on phytoplankton communities in Australian waters. Units and taxonomy have been standardised, obviously erroneous data removed, and all metadata included. We have lodged this dataset with the Australian Ocean Data Network (http://portal.aodn.org.au/) allowing public access. The Australian Phytoplankton Database will be invaluable for global change studies, as it allows analysis of ecological indicators of climate change and eutrophication (e.g., changes in distribution; diatom:dinoflagellate ratios). In addition, the standardised conversion of abundance records to biomass provides modellers with quantifiable data to initialise and validate ecosystem models of lower marine trophic levels.

Identifying habitats at risk: simple models can reveal complex ecosystem dynamics.

The relationship between ecological impact and ecosystem structure is often strongly nonlinear, so that small increases in impact levels can cause a disproportionately large response in ecosystem structure. Nonlinear ecosystem responses can be difficult to predict because locally relevant data sets can be difficult or impossible to obtain. Bayesian networks (BN) are an emerging tool that can help managers to define ecosystem relationships using a range of data types from comprehensive quantitative data sets to expert opinion. We show how a simple BN can reveal nonlinear dynamics in seagrass ecosystems using ecological relationships sourced from the literature. We first developed a conceptual diagram by cataloguing the ecological responses of seagrasses to a range of drivers and impacts. We used the conceptual diagram to develop a BN populated with values sourced from published studies. We then applied the BN to show that the amount of initial seagrass biomass has a mitigating effect on the level of impact a meadow can withstand without loss, and that meadow recovery can often require disproportionately large improvements in impact levels. This mitigating effect resulted in the middle ranges of impact levels having a wide likelihood of seagrass presence, a situation known as bistability. Finally, we applied the model in a case study to identify the risk of loss and the likelihood of recovery for the conservation and management of seagrass meadows in Moreton Bay, Queensland, Australia. We used the model to predict the likelihood of bistability in 23 locations in the Bay. The model predicted bistability in seven locations, most of which have experienced seagrass loss at some stage in the past 25 years providing essential information for potential future restoration efforts. Our results demonstrate the capacity of simple, flexible modeling tools to facilitate collation and synthesis of disparate information. This approach can be adopted in the initial stages of conservation programs as a low-cost and relatively straightforward way to provide preliminary assessments of.nonlinear dynamics in ecosystems.

Recovery of a subtropical rocky shore is not yet complete, four years after a moderate sized oil spill.

Little is known about the recovery trajectory from small to moderate spills (<1000 t), particularly in the sub-tropics. On 11 March 2009 the MV Pacific Adventurer spilt 270 t of bunker fuel oil 13 km off Moreton Island, Australia, impacting wetlands, sandy beaches and rocky shores. This study examines the recovery of the rocky shore community four years after the spill. Results indicate that recovery on Moreton Island is taking longer than the 3-4 years suggested by the literature. The upper shore is recovering faster than the mid shore and is nearly recovered while the mid shore is still in the recovery process. These results indicate that small to moderate sized spills can have environmental impacts on par with much larger spills and emphasizes the need for a clear definition of a recovery endpoint. Long term studies are required to gain a full understanding of trajectories of recovery after oil spill impacts.

Marine reserves help coastal ecosystems cope with extreme weather.

Natural ecosystems have experienced widespread degradation due to human activities. Consequently, enhancing resilience has become a primary objective for conservation. Nature reserves are a favored management tool, but we need clearer empirical tests of whether they can impart resilience. Catastrophic flooding in early 2011 impacted coastal ecosystems across eastern Australia. We demonstrate that marine reserves enhanced the capacity of coral reefs to withstand flood impacts. Reserve reefs resisted the impact of perturbation, whilst fished reefs did not. Changes on fished reefs were correlated with the magnitude of flood impact, whereas variation on reserve reefs was related to ecological variables. Herbivory and coral recruitment are critical ecological processes that underpin reef resilience, and were greater in reserves and further enhanced on reserve reefs near mangroves. The capacity of reserves to mitigate external disturbances and promote ecological resilience will be critical to resisting an increased frequency of climate-related disturbance.

Recovery of a freshwater wetland from chemical contamination after an oil spill.

In March 2009, a cargo ship spilled 250 tons of heavy fuel oil off the Queensland coast of Australia. The pristine National Park Moreton Island, seven nautical miles to the east of the spill site, was most affected by the oil slick. Contamination of the island's shoreline was widespread, with freshwater wetlands particularly slow to recover as clean-up needed to be carefully managed to avoid damage to this sensitive ecosystem. During the clean-up process on Moreton Island a monitoring program was initiated using traditional chemical analysis in combination with bioanalytical techniques to assess the extent and variability in contamination at sites on the shoreline and freshwater wetlands. Water accommodated fractions (WAF) of oil residues from samples taken directly after the spill on the shoreline showed the same level of toxic potency as samples from the wetland while baseline-toxicity equivalent concentrations (baseline-TEQ) and 2,3,7,8-tetrachlorodibenzodioxin equivalent concentrations (TCDDEQ) were much lower in oil collected from the sandy beach. The umuC assay for genotoxicity and the E-SCREEN assay for estrogenic effects indicated the extracts were not genotoxic or estrogenic. PAH concentrations and toxicity in grab water samples were below detectable levels, however, extracts from time integrated silicone passive samplers deployed for several weeks at the contaminated sites gave measurable responses in the bioassays with TCDDEQ levels increased relative to the control site. The low levels of baseline-TEQ and TCDDEQ present after 8 months had further decreased 6 months later indicating satisfactory recovery of this pristine ecosystem after an oil spill.