Spencer, M., Birchenough, S.N.R., Mieszkowska, N., Robinson, L.A., Simpson, S.D., Burrows, M.T., Capasso, E., Cleall-Harding, P., Crummy, J., Duck, C., Eloire, D., Frost, M., Hall, A.J., Hawkins, S.J., Johns, D.G., Sims, D.W., Smyth, T.J. and Frid, C.L.J. (2011) Temporal change in UK marine communities: trends or regime shifts? Marine Ecology, 32 (Suppl. S1), 10-24. (doi:10.1111/j.1439-0485.2010.00422.x).
Abstract
A regime shift is a large, sudden, and long-lasting change in the dynamics of an ecosystem, affecting multiple trophic levels. There are a growing number of papers that report regime shifts in marine ecosystems. However, the evidence for regime shifts is equivocal, because the methods used to detect them are not yet well developed. We have collated over 300 biological time series from seven marine regions around the UK, covering the ecosystem from phytoplankton to marine mammals. Each time series consists of annual measures of abundance for a single group of organisms over several decades. We summarised the data for each region using the first principal component, weighting either each time series or each biological component (e.g. plankton, fish, benthos) equally. We then searched for regime shifts using Rodionov’s regime shift detection (RSD) method, which found regime shifts in the first principal component for all seven marine regions. However, there are consistent temporal trends in the data for six of the seven regions. Such trends violate the assumptions of RSD. Thus, the regime shifts detected by RSD in six of the seven regions are likely to be artefacts caused by temporal trends. We are therefore developing more appropriate time series models for both single populations and whole communities that will explicitly model temporal trends and should increase our ability to detect true regime shift events.
More information
Identifiers
Catalogue record
Export record
Altmetrics
Contributors
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.