USING BIVALVE LIVE-DEAD MISMATCH TO EVALUATE THE FOOTPRINT OF MUNICIPAL WASTEWATER IN MARINE WATERS, VICTORIA, BRITISH COLUMBIA

Contrasts in the species composition of living communities and naturally time-averaged death assemblages, extracted from the same benthic samples, have potential to detect changes in ecosystem state driven by human activities. Not all ‘stressed’ settings exhibit live-dead discordance: not all stresses cause ‘strain’, pre-stress death assemblages can be deeply buried where stress is long-standing, and high sedimentation or shell loss can minimize time-averaging. Here, we use bivalve live-dead analysis to assess the footprint of a wastewater outfall, active since the 1970s, on cold-temperate shelf communities (50 m). Regional effects have been judged minimal given naturally high turbulence, oxygenation, and productivity in the Juan de Fuca Strait, but the limited level of wastewater treatment remains politically charged. Municipal agency data on living benthic communities sampled in 2005-2014 were pooled to establish average bivalve species composition per site; sites within 800m are categorized as near-, mid-, and far-field habitats based on sediment metal content, contrasted with a reference area 8 km away. Sieve residues from 2014 samples were picked for dead individuals, which had to retain ≥ one half of the hingeline.

We find (a) between-habitat variation in species and functional-group composition in living bivalves consistent with nutrification, signals of which are observed around the outfall in other routine sediment monitoring endpoints. (b) Dead shells are ~as abundant and diverse as living individuals and can be badly corroded, suggesting high post-mortem loss. However, dead shells detect both (c) key spatial variation among living chemosymbiotic bivalves -- pollution-tolerant lucinids decrease relative to thyasirids with distance -- and (d) two sites with unusual communities, indicating little spatial blurring. (e) Live-dead discordance nonetheless exists in relative abundances, with mismatch greater closer to the outfall: detritus-based functional groups are less abundant dead than alive (8-36% dead v. 40-70% living), and epi- and infaunal suspension-feeders are more abundant dead, including in the reference area (40% of dead, 20% of living). Shell age-dating is needed to confidently implicate late 20^th century stresses, but our results suggest a nutrient footprint ≥1-km.