Assessment of the effects of the inclusion of poor quality sediment samples on spatial pre

Over 7,000 samples were excluded from the Australian Marine Samples (MARS) database due to quality control, primarily dredged samples. This dataset from the Australian Ocean Data Network contains 14,204 seabed samples used to assess how including poor-quality dredged samples affects the accuracy of spatial predictions for mud content. The analysis focuses on two regions, the Southwest and Petrel, using methods like Inverse Distance Weighting and Ordinary Kriging.

Use Cases

• Assessing the impact of dredged sample inclusion on seabed mud content prediction accuracy based on the described cross-validation analysis.
• Comparing spatial interpolation methods like Inverse Distance Weighting and Ordinary Kriging for marine sediment mapping.
• Evaluating data quality control protocols for large marine sample databases like MARS.
• Studying regional variations in sediment composition across the Australian Exclusive Economic Zone.

Strengths

• Based on a substantial database of 14,204 seabed samples.
• Provides a detailed methodological comparison of spatial interpolation techniques (IDW and OK).
• Quantifies prediction accuracy using relative mean absolute error (RMAE) with specific percentage differences reported.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the specific analysis dataset is unknown, which may limit suitability assessment.
• Data coverage is described as uneven across the Australian Exclusive Economic Zone.

Provenance

Source

Australian Ocean Data Network

Collection Method

Samples collected and stored in the Australian Marine Samples (MARS) database, with analysis involving spatial interpolation and statistical testing.

Freshness

Last updated 2026-06-05 07:17:05.825170; freshness should be verified.

Geography

Australian Exclusive Economic Zone, specifically the Southwest and Petrel regions.