This paper describes a novel statistical approach to derive ecologically relevant sediment quality guidelines (SQGs) from field data using a nonparametric empirical Bayesian method (NEBM). We made use of the Norwegian Oil Industrial Association database and extracted concurrently obtained data on species density and contaminant levels in sediment samples collected between 1996 and 2001. In brief, effect concentrations (ECs) of each installation (i.e., oil platform) at a given reduction in species density were firstly derived by fitting a logistic-type regression function to the relationship between the species density and the corresponding concentration of a chemical of concern. The estimated ECs were further improved by the NEBM which incorporated information from other installations. The distribution of these improved ECs from all installations was determined nonparametrically by the kernel method, and then used to determine the hazardous concentration (HC) which can be directly linked to the species loss (or the species being protected) in the sediment. This method also enables an accurate estimation of the lower confidence limit of the HC, even when the number of observations was small. To illustrate the effectiveness of this novel technique, barium, cadmium, chromium, copper, mercury, lead, tetrahydrocannabinol, and zinc were chosen as example contaminants. This novel approach can generate ecologically sound SQGs for environmental risk assessment and cost-effectiveness analysis in sediment remediation or mud disposal projects, since sediment quality is closely linked to species density. Copyright © 2013 Springer-Verlag Berlin Heidelberg.
CitationLui, G. C. S., Li, W. K., Bjørgesæter, A., & Leung, K. M. Y. (2014). Deriving field-based sediment quality guidelines from the relationship between species density and contaminant level using a novel nonparametric empirical Bayesian approach. Environmental Science and Pollution Research, 21(1), 177-192. doi: 10.1007/s11356-013-1889-1
- Sediment quality guidelines
- Species sensitivity distribution
- Ecological modeling
- Empirical Bayesian methods
- Benthic biodiversity