Till startsida
To content Read more about how we use cookies on gu.se

Modelling benthic habitats

The distributions of benthic species and habitats vary typically at multiple spatial and temporal scales. In PREHAB we review and incorporate existing knowledge into empirical models to predict the spatial distribution of benthic habitats in a set of case-study areas. Existing information involves data on environmental factors such as bathymetry e.g. depth and slope, wave-exposure and biological variables such as vegetation cover, species composition of rocky and sediment assemblages and fish recruitment areas.

There is a wide array of statistical techniques available for construction of such predictive models. In PREHAB one all-embracing aim is to compare the performance of techniques in order to arrive at recommendations for cost-efficient and precise modelling practices.

Mapping procedure

Predictive mapping involves four main steps: (1) Sampling of biological data using appropriate sampling methods (e.g. video or grabs) at a limited number of sites (2) compilation of full-coverage physical data (e.g. depth, exposure, substrate) from GIS-sources; (3) development of statistical model from data using appropriate statistical methods (e.g. generalised linear models or tree-based methods) and (4) construction of full-coverage GIS-layer of biological features using the model and full-coverage physical data.

Preliminary results

PREHABs´ research on predictive modeling and mapping is based on conclusions from four case-study areas. Using existing data in these areas we examine crucial issues like:

  • What can be modelled?
  • Which type of information can be used for modeling?
  • Which statistical tools can be used for modeling?


Results 2010

We need maps!

In order to incorporate ecology into planning we need spatial information - maps - about the distribution of ecological properties. Because of the methods used and the large costs associated with biological sampling benthic assemblages, construction of detailed large-scale maps from direct measurements is usually not possible. Sampling coupled to predictive models is usually a more realistic approach.


Other marine modelling projects


© University of Gothenburg, Sweden Box 100, S-405 30 Gothenburg
Phone +46 31-786 0000, About the website

The University of Gothenburg uses cookies to provide you with the best possible user experience. By continuing on this website, you approve of our use of cookies.  What are cookies?