Summary
Chemical contamination resulting from human activities has been identified in nearshore marine environments in the vicinity of Antarctic Research Stations. To better manage and remediate contaminated sites, and to assess the risks of ongoing activities, Water Quality Guidelines incorporating sensitivities of Antarctic species and specific characteristics of their environment are required. Previous toxicity tests with Antarctic invertebrates have shown that responses to metal exposure are far slower than in traditional tests. The slow metabolism of Antarctic organisms may result in longer response times, yet does not mean these species are less sensitive to contaminants than their tropical or temperate counterparts. We investigated the effects of five metals on behaviour and survival of the Antarctic amphipod Orchomenella pinguides over 30 days. Multiple observations were made through time to assess mortality rate response. A new statistical model that combines the bioassay model with a survival time model was developed as an alternative to traditional Probit Analysis for individual time endpoints which are treated independently. Interval?censored survival times were modeled using a generalised additive model (GAM) with multiplicative effects for concentration level. The time period by concentration level interaction was included as a random effect term with this mixed model version (GAMM). This approach is advantageous as it smoothes through noisy periodic mortality data, providing a more powerful tool by using all data simultaneously and estimating trends across multiple observation periods. It also improves models of uncertainty in estimates of lethal concentrations. Response times of amphipods varied between metals, with behavioural changes and mortality occurring at a faster rate in copper than in other metals (cadmium, lead, zinc or nickel). Results from comparisons of both approaches indicate that these modified statistical methods provide improved estimates of sensitivity of Antarctic species which will inform more accurate site specific Water Quality Guidelines.