Methods for evaluating changes in benthic communities

Environmental impact factors, or EIF, is a method developed for the petroleum industry to assess the environmental impacts from drilling discharges. Analysis of changes in benthic communities is one of the key assessment tools used in the EIF approach. A variety of statistical tools are available to detect changes in benthic communities. In order to optimize the EIF methodology for assessing impacts from drill cuttings discharges, Akvaplan-niva carried out an inter-comparison assessment of these methods. The goal of the project was to identify methods that are best suited to quantify the relationship between cause and effect.

The assessment showed that the most suitable approach was the multivariate statistical method known as Community Disturbance Index (CDI); a method based on establishing a reference area to identify the natural variations in the area of study. The CDI method fulfils the core criteria for successfully identifying changes in benthic communities related to disturbances, namely it is highly sensitive and objective, applicable for modelling and impact assessments, and recognized and accepted by both regulatory authorities and the scientific community.

Financed by the Oljeindustriens Landsforening

Figure: Available methods for benthic community disturbance analysis as part of the EIF management tool.

Tool	Advantages	Drawbacks	Suitability in EIF
Chemical analyses
LSC (Limit of Significant Contamination)	Calculates specific value for contamination limits.	Dependent of the selection of stations included in the calculation. Does not give the status for the biological condition.	Unsuitable alone
Univariate methods
Shannon-Wiener H’	Internationally well-known and used statistic. Calculates a specific value relating to disturbance.	Not accepted as a measurement of effects of pollution.	Unsuitable alone
Hurlbert’s ES₁₀₀	Internationally well-known and used statistic. Calculates a specific value relating to disturbance.	As above.	Unsuitable alone
Pielou’s J	Internationally well-known and used statistic. Calculates a specific value relating to disturbance.	As above.	Unsuitable alone
Graphical/distributional methods
Log-normal plots	Internationally well-known and used statistic.	Might indicate disturbance, but needs detailed information on dominant species. Difficult to quantify.	Unsuitable alone
Multivariate methods
Cluster analysis	An internationally well-known and well-used method for many years.	Shows faunal differences between stations, but gives no information on why such differences occur and what the differences are.	Unsuitable alone. Could possibly be used together with other methods.
MDS	An internationally well-known and well-used method for many years.	As for the cluster analysis. However, combined with the results from other analyses the reason for the gradient can be identified. The scale of the method is relative, which makes it difficult to assess the significance of the observed differences between samples in the plot.	Unsuitable alone. Could possibly be used together with other methods
CA	An internationally known and used method.	As for the Cluster analysis and MDS.	Unsuitable alone
CDI	Combines biological and environmental variables. Calculates a specific value relating to disturbance.	The method has until recently not been available. Not peer-reviewed nor internationally known or used.	Best developed for EIF so far. Needs further evaluation and practical use by external groups
ANOSIM	Calculates whether group of samples are significantly different from each other.	Based on relative differences between samples, as MDS.	Unsuitable alone
Relationship between biological and environmental variables
CCA	Combines biological and environmental variables and shows the correlations between significant environmental factors and the faunal distribution. Internationally well-known and used.	Does not have a scale relating to disturbance.	Could possibly be used. Needs further development
BIO-ENV	Combines biological and environmental variables. Internationally well-known and used.	Does not have a scale relating to disturbance. Does not indicate how particular stations or species are related to disturbance.	Could possibly be used. Needs further development
Qualitative methods
Top ten and indicator species	Gives specific information on the dominant species. Can be quantified.	Has not been scaled earlier. Needs to be adjusted for each specific field.	Unsuitable alone. Could possibly be used together with other methods.
Faunal groups	Evaluate the results from all analyses carried out (univariate-, multivariate- and correlation analyses). Can be quantified.	Subjective and dependent on the person/institution carrying out the evaluation.	Unsuitable alone. Could possibly be used together with other methods.

• Previous page