Abstract
Understanding the influence of macroinvertebrates on ecosystem function often relies on experimental defaunation with methods that remove fauna through minimal sample disturbance. Defaunation is challenging and can lead to confounding effects and/or loss of empirical information when unsuccessful. We evaluated the ability of a deoxygenation treatment to remove macroinvertebrates from sediment cores collected in 2 regions of a microtidal estuary. Only 1 of 16 cores was fully defaunated following 3 deoxygenation cycles. To counteract confounding effects of partial defaunation, we quantified the biomass remaining in each core and used these data as a covariate in statistical models. The unremoved biomass had, in some cases, significant effects on alkalinity fluxes, with positive linear relationships evident, and net phosphate fluxes. The community in the upper estuary that regularly experiences hypoxia exhibited stronger sediment emergence responses (82-100%). The remaining fauna were spread equally among annelids, molluscs and arthropods in abundance, although arthropods dominated the biomass. In contrast, fewer macroinvertebrates emerged from sediments from the lower estuary (47-89%), with most of the remaining biomass and abundance being annelids and molluscs. These findings suggest that estuarine taxa have divergent responses to hypoxia and that regional communities are variably prone to eradication of sensitive taxa. Our study shows how the use of defaunation by deoxygenation can create systematic bias, particularly when comparing areas with disparate in situ oxygen regimes, and provides a way to quantitively account for partial defaunation without sacrificing statistical power or using overly destructive methods.
