Multispectral multibeam backscatter response of heterogeneous rhodolith beds

Area of Science:

• Marine geology
• Seafloor mapping
• Biogenic habitat characterization

Background:

• Acoustic backscatter is crucial for detailed seafloor mapping, yet distinguishing similar substrates like rhodoliths remains challenging.
• Rhodolith beds, vital biogenic habitats, are often broadly classified despite variable nodule coverage.
• Multispectral acoustic backscatter shows promise for enhanced thematic seafloor mapping over single-frequency systems.

Purpose of the Study:

• To characterize and map rhodolith beds using multispectral acoustic backscatter and underwater imagery.
• To assess the effectiveness of multispectral backscatter in differentiating rhodolith coverage proportions.
• To evaluate the predictive capability of acoustic backscatter for rhodolith presence and absence.

Main Methods:

• Utilized multispectral multibeam backscatter and underwater imagery for seafloor characterization.
• Employed a support vector machine classifier to analyze multifrequency backscatter mosaics.
• Correlated acoustic data with rhodolith classes identified from in-situ imagery.

Main Results:

• Multispectral acoustic backscatter effectively mapped varying proportions of rhodolith coverage.
• The lowest acoustic frequency best distinguished different rhodolith coverage levels.
• Higher acoustic frequencies accurately predicted the presence and absence of rhodoliths.

Conclusions:

• Multispectral acoustic backscatter significantly improves the thematic mapping of seafloor substrates like rhodolith beds.
• This method provides valuable data for understanding and managing ecologically important biogenic habitats.
• Combining multiple acoustic frequencies offers a robust approach for detailed seafloor characterization.