- The inherent complexity of the environment is such that attempts to model it must operate under simplifications and assumptions. Considering predictions from alternative models, with a range of assumptions and data requirements, therefore provides a more robust approach.
- The intractability and uncertainty resulting from a suite of predictions may hinder the application of science in policy, where a single prediction with little ambiguity or uncertainty would be most desirable. Few studies modelling species' distributions attempt to present multi-model outputs in a format most useful to the non-modelling community, and none of these have done so for the marine environment.
- The problem of uncertainty is particularly prevalent in predicting the distribution of invasive alien species under climate change. As invasive alien species are one of the main drivers of biodiversity loss and may incur significant economic costs, the benefit of applying predictions to highlight areas of possible establishment and inform policy and management may be large.
- An ensemble prediction is used to assess the distribution of suitable habitat for the Pacific oyster, Crassostrea gigas, in UK waters both currently and in the future. The ensemble incorporates predictions from three species distribution models, using data from two global climate models. A method is developed highlighting the agreement of the ensemble, further applying threshold values to retain information from constituent predictions in the final map of agreement.
- Ensemble predictions made here suggest that Pacific oyster will experience an opening of suitable habitat in northern UK waters, reaching the Faroe Islands and the eastern Norwegian Sea by 2050. Habitat suitability will increase with warming temperatures in the English Channel and Central North Sea for this species. The approaches applied here can be incorporated into risk assessment frameworks for invasive species, as stipulated in the Convention on Biological Diversity.
Copyright © 2013 John Wiley & Sons, Ltd.