Are conventional statistical techniques exhaustive for defining metal background concentrations in harbour sediments? A case study: The Coastal Area of Bari (Southeast Italy)

[Display omitted] •Metal background concentrations in South-East Italy sediments were calculated.•Three conventional statistical techniques were used and compared to each other.•The conventional techniques led to the same range of heavy metal background values.•Mineralogical and multivariate statistical analyses revealed new specific aspects.•Heterogeneously constituted areas need deeper investigations. Sediment contamination by metals poses significant risks to coastal ecosystems and is considered to be problematic for dredging operations. The determination of the background values of metal and metalloid distribution based on site-specific variability is fundamental in assessing pollution levels in harbour sediments. The novelty of the present work consists of addressing the scope and limitation of analysing port sediments through the use of conventional statistical techniques (such as: linear regression analysis, construction of cumulative frequency curves and the iterative 2σ technique), that are commonly employed for assessing Regional Geochemical Background (RGB) values in coastal sediments. This study ascertained that although the tout court use of such techniques in determining the RGB values in harbour sediments seems appropriate (the chemical–physical parameters of port sediments fit well with statistical equations), it should nevertheless be avoided because it may be misleading and can mask key aspects of the study area that can only be revealed by further investigations, such as mineralogical and multivariate statistical analyses.