AUTHOR=Linklater Michelle , Morris Bradley D. , Hanslow David J. TITLE=Classification of seabed landforms on continental and island shelves JOURNAL=Frontiers in Marine Science VOLUME=10 YEAR=2023 URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2023.1258556 DOI=10.3389/fmars.2023.1258556 ISSN=2296-7745 ABSTRACT=

The increasing availability and quality of high-resolution bathymetry data has led to a growing need for automated classification approaches to extract seabed features and better understand our ever-changing and complex seascapes. Here we present a new set of GIS tools designed to classify seabed landforms on continental and island shelf settings. The classification approach utilises bathymetry data and its derivatives of slope, ruggedness and bathymetric position index to delineate key components of the seabed surface. The user is guided through a series of steps to break down the seabed surface into components termed ‘surface elements’ (e.g. smooth, rugose, slope areas), which are subsequently grouped into prominent seabed features termed ‘seabed landforms’ (e.g. reefs, channels, scarps). Manual review and editing are incorporated into the workflow, striking a balance between automation and expert manual interpretation. We present the toolset using examples from the statewide marine lidar dataset from New South Wales, Australia, and explore tool settings using bathymetric data representing different data sources (multibeam and marine lidar), environmental seascapes, data resolutions (2, 5, 10 and 20 m cell size) and data preparation treatments (with and without data smoothing). The GIS toolset presented offers an effective and flexible method to extract key features from high-resolution shelf bathymetry data. Such mapping provides fundamental baseline data for vast applications within marine planning, research and management.