Publications: Peer-reviewed journal articles (by staff)
Assessing ecological community health in coastal estuarine systems impacted by multiple stressors
Ellis JI, Hewitt JE, Clark D, Taiapa C, Patterson M, Sinner J, Hardy D, Thrush SF 2015. Assessing ecological community health in coastal estuarine systems impacted by multiple stressors. Journal of Experimental Marine Biology and Ecology 473: 176-187.
DOI link here.
Increasing population pressure, urbanisation of the coastal zone and nutrient and sediment run-off from agriculture and forestry has increased the number of large-scale and chronic impacts affecting coastal and estuarine systems. The need to assess cumulative impacts is a major motivation for the current desire of managers and ecologists to define ecosystem “health” and “stress”. A number of univariate metrics have been proposed to monitor health, including indicator species, indicator ratios and diversity or contaminant metrics. Alternatively, multivariate methods can be used to test for changes in community structure due to stress. In this study we developed multivariate models using statistical ordination techniques to identify key stressors affecting the ‘health’ of estuarine macrofaunal communities. Macrofaunal and associated environmental samples were collected across 75 sites from within Tauranga Harbour, a large estuary located on New Zealand's North Island. The harbour receives discharges from urbanised, industrial, agricultural and horticultural catchments. Distance-based linear modelling identified sediments, nutrients and heavy metals as key ‘stressors’ affecting the ecology of the harbour. Therefore, three multivariate models were developed based on the variability in community composition using canonical analysis of principal coordinates (CAP). The multivariate models were found to be more sensitive to changing environmental health than simple univariate measures (abundance, species richness, evenness and Shannon–Wiener diversity) along an anthropogenic stress gradient. This multivariate approach can be used as a management or monitoring tool where sites are repeatedly sampled over time and tracked to determine whether the communities are moving towards a more healthy or unhealthy state. Ultimately, such statistical models provide a tool to forecast the distribution and abundance of species associated with habitat change and should enable long-term degradative change from multiple disturbances to be assessed.