Abstract


DIGITIZING THE CORAL REEF: SURVEYING THE BIODIVERSITY OF CORAL REEF BENTHOS USING UNDERWATER HYPERSPECTRAL IMAGING

Coral reefs are characterized by a high degree of spatial complexity and biodiversity. Rapid large-scale surveys at a fine spatial resolution and automation of substrate recognition will help to quantify the spatial structure of the reef's biodiversity. This is a pre-requisite to assess changes, both natural and anthropogenic, in the ecological status of coral reefs. We describe a novel method to survey and map the biodiversity of coral reef benthos using a diver-operable hyperspectral imager called hyperdiver. The hyperdiver survey technique captures spectral and visual images, topography, water chemistry and light field data from survey transects. The technique combines the information derived from high-resolution spectra and imagery to feed a machine learning analysis, which can be used to create classified reef maps down to the taxonomic level of genera. Comparison of machine classification against expert analysis of taxa shows very good congruence. Consequently, the hyperdiver technique allows us to rapidly visualize the biodiversity of the reef benthos from the centimeter scale upwards, along with profiling the bedform rugosity and light climate. We used hyperdiver to survey unique Pacific reefs near natural CO2 vents and reefs nearby unaffected by elevated CO2 levels. The survey covered several kilometers, and the data was used to create classified reef maps with 1-2 cm spatial resolution. We visualized and analyzed the strong effects of ocean acidification on the composition and rugosity of the benthic assemblage from the centimeter to reef scales.

Authors

Chennu, A., Max Planck Institute for Marine Microbiology, Germany, achennu@mpi-bremen.de

den Haan, J., Max Planck Institute for Marine Microbiology, Germany, jhaan@mpi-bremen.de

Fabricius, K., Australian Institute of Marine Science, Australia, K.Fabricius@aims.gov.au

de Beer, D., Max Planck Institute for Marine Microbiology, Germany, dbeer@mpi-bremen.de

Details

Oral presentation

Session #:53
Date: 06/24/2016
Time: 15:30
Location: 312

Presentation is given by student: No