Exploiting salmon farm benthic enrichment gradients to evaluate the regional performance of biotic indices and environmental indicators

This study evaluates five benthic indicators (total abundance, number of taxa, redox potential, total free sulfides, total organic matter) and ten biotic indices (Margalef's d, Peilou's J’, Shannon H’, AMBI, M-AMBI, MEDOCC, BENTIX, BOPA, ITI, BQI), to identify those that best define organic enrichment gradients under different flow regimes. Performance was measured against Enrichment Stage (ES), a continuous variable characterising the full range of sediment conditions (natural to azoic). None of the 15 metrics were able to consistently discriminate over the full enrichment gradient for both flow environments. The most versatile indices were BQI > M-AMBI > AMBI > Log( N ) > BENTIX. Of these, M-AMBI best catered for different flow environments, while the BQI was the most effective under highly enriched conditions. Under strong enrichment, i.e. when macrofauna abundance is in decline, changes in redox, sulfides, number of taxa and abundance were reasonably clear. However, the more complex biotic indices were relatively insensitive at this level, highlighting a limited applicability beyond the ‘peak of opportunists’ (PO). Conversely, in high flow regimes, some of the biological indicators were relatively sensitive to low-to-moderate levels of enrichment that were not well discerned by the physico-chemical variables. A useful subset of variables for assessing enrichment status is recommended, comprising two of the best performing biotic indices that are based on alternative/independent classification schemes (i.e. EG's and ES50 0.05 ), total abundance, to aid in discerning PO, and a geochemical variable (redox or S 2− ). Inconsistencies between metrics were found to be more significant than the variability surrounding the predictive capacity of individual indicators, and as a result there is a risk of ES misclassification where only a single index is used. Whilst there is a recognised need to use combinations of indicators, this study also stresses the importance of focusing on a few regionally validated measures and down-weighting the importance placed on any that are not. Additionally, although using a combination of different indicators may produce a ‘safe’ average result, it may be inefficient, and the averaging effect has the potential to mask extreme conditions. Hence, there remains a need for expert judgement to select and appropriately weight indicator variables, to identify any erroneous results, and to reliably assess ecological quality status.