Predictive habitat modelling

This is a widely used, automated process for classifying benthic habitats. It utilises a variety of high-resolution physical variables identified as proxies for habitats and the composition of species and communities associated with those habitats.

Why work with Ocean Ecology?

We offer expertise

Our GIS specialists use automated predictive mapping and expert interpretation of acoustic and ground-truthing data to generate predictive habitat maps.

They apply supervised learning to predict habitat classes, guided by cross-scale seabed morphological features. Bathymetric data, high resolution SSS mosaic and broadscale current speed and wave data help compute the model while the results of seabed imagery analysis and sediments characteristics are used to ground-truth and train the model.

Ever expanding experience

The number of applications for predictive modelling is constantly growing. We have used the technique to map large areas of the seabed including many complex features that would otherwise have been very time consuming to map manually.