Spatio-temporal dynamics of temperate marine systems; species, communities and habitat.

Abstract

Global human populations rely on the oceans as a source of food and other resources. In order to maintain this resource, information about seabed habitat is essential to scientists and marine managers to provide a baseline against which to assess future change. These data are often supplied as maps. For this information to be useful, it must be based on sound scientific knowledge. Many marine mapping studies compare variable selection in terms of the environmental data that go into habitat maps. Similarly, many studies have compared classification algorithms. However, these two are rarely considered together. This study compares datasets along with two approaches to marine habitat map production. At the same time, this study investigates the spatial scales at which diversity and species associations can be monitored. A high resolution hydrodynamic model is produced for the primary study area, and this is integrated with predictive models. Temporal stability of communities is assessed, and subsequently a single species detection study is conducted. Unsupervised clustering routines are shown to be just as ecologically meaningful as supervised classifications in areas of hard substrata. Patterns in species association and diversity are most pronounced across meso-scales; future species distribution modelling studies should take these scales into account. Temporal stability of benthic community structure is high as indicated by numerical models, however the small amount of error indicated by numerical models propagates through analyses to create a large difference in output classified maps. Hydrodynamics becomes the second most important variable in describing species distributions, and reveal previously unknown behaviour of a temperate marine fish of high conservation status. Detectability is highlighted as an important consideration for future marine ecology studies.

Date of Award	31 May 2018
Original language	English
Sponsors	Department of Education and Learning (DEL)
Supervisor	Rory Quinn (Supervisor) & Chris Mc Gonigle (Supervisor)