Abstract
The aim of this project is develop empirical, predictive models of benthic diversity in deep marine habitats. This vill be done by testing hypotheses about patterns of (a) spatial variability, (b) associations among benthic species and (c) relationships among assemblages and physical properties of the habitat. Data on species richness, composition and abundance will be obtained by quantitative analysis of existing video-recordings and high-resolution (10 x 10 m) bathymetric data from multi-beam sonar scans will be used to obtain estimates of depth, slope, topographic complexity, substratum porosity, etc. Explicit tests of the above hypotheses will be tested using a suite of uni- and multivariate techniques for analysis of spatial patterns and association among variables. The existence of any correlative pattern will determine the exact structure and scope of predictive models. The predictive power and precision of models will be tested on independent data. One fundamental result in contemporary ecology is the realisation that spatial (and temporal) variability is scale-dependent and that the importance of ecological processes differ among scales. This has important consequences for the study of spatial patterns and for the development of mechanistic as well as purely predictive models. Combining this knowledge with the recent access to sophisticated technologies for the measurement of marine biodiversity is consitutes an important task and an interesting challenge.
Sweden