Seabed Mud Content Simulation for Australian Margin Interpolation

A simulation experiment from Geoscience Australia compares statistical and mathematical techniques for predicting seabed mud content. The study evaluates factors like sample density, search neighborhoods, and secondary variables including bathymetry and distance-to-coast. It identifies a novel combined method, random forest and ordinary kriging (RKrf), which reduced relative mean absolute error by up to 17% compared to a control.

Use Cases

• Compare spatial interpolation methods like inverse distance squared, random forest, and ordinary kriging using simulated seabed mud content data.
• Analyze the effect of sample density and search neighborhood size on prediction accuracy metrics such as mean absolute error.
• Model seabed mud content using secondary variables bathymetry, distance-to-coast, and slope as covariates.
• Assess prediction accuracy across different Australian regions (north, northeast, southwest) using relative mean absolute error (RMAE).
• Apply the identified robust method (RKrf) to model other physical properties for marine biodiversity prediction.

Strengths

• Study evaluates five key factors affecting interpolation accuracy: regions, methods, sample densities, searching neighborhoods, and sample stratification.
• Validation uses ten-fold cross-validation and multiple accuracy metrics including mean absolute error and root mean square error.
• Identified combined method (RKrf) achieved a relative mean absolute error up to 17% lower than the control method.

Limitations

• Dataset is a simulation experiment derived from the Geoscience Australian Marine Samples database, not direct observational measurements.
• The raw description discusses data noise and required quality control criteria, indicating potential preprocessing needs for the underlying sample data.
• Specific sample size, row count, and column details for the simulation data are not provided.

Provenance

Source

Geoscience Australia Data

Collection Method

Simulation experiment using samples from the Geoscience Australian Marine Samples database, with data quality control applied.

Time Range

null

Freshness

Last updated March 2026.

Geography

Australian marine margin, with analysis across specific regions (north, northeast, southwest).