Comparison of sampling methods for deep‐sea infauna

Abstract

Sampling methods for benthic meiofauna and macrofauna assessments on the northern Gulf of Mexico continental slope and deep sea were compared. For meiofauna, a core with an inner diameter of 5.1 cm is recommended for yielding an appropriate sample size. Meiofauna are concentrated in the uppermost 2 cm sediment layer, so the top 3 cm are sufficient to sample. Macrofauna penetrate deeper and the top 10 cm are sufficient. Smaller sieves capture more organisms so 45 μm for meiofauna, and 300 μm for macrofauna, is recommended. On average, 88% of meiofauna were extracted in the Ludox fraction compared to the total of both Ludox and the sediment pellet. Box corers and multiple corers were compared for estimating macrofauna and meiofauna metrics. Multicorers are recommended for quantitative assessments, but box corers are useful for qualitative studies that require capturing more diversity. Box cores underestimate macrofauna abundance by 2.9 times. While the larger box core captures more species resulting in higher diversity estimates, it is low relative to the 24 times larger area sampled. The multicorer preserves vertical distribution. Because meiofauna are sampled from subcores, there is little difference between the two devices for estimating meiofauna metrics. Replicate multicore samples (i.e., deployments) do not add substantially to our understanding of the variance of species richness or abundance, thus to describe the spatial footprint of macrofauna community structure, it is recommended that resources should be used to sample more stations over a larger area rather than multiple replicates at fewer stations.

Macrobenthic infaunal communities associated with deep‐sea hydrocarbon seeps in the northern Gulf of Mexico

“…Montagna et al () compared benthic communities collected in the deep Gulf of Mexico at the same place and time by both a multicorer and boxcorer. They found that the boxcorer underestimated macrofaunal abundance by 3x although the multicorer collected 60% fewer taxa than the boxcorer.…”

“…Bow waves from the boxcorer may wash away small, surface dwelling animals during collections (Hulings & Gray, ). The washing and sieving of samples on the boat deck during boxcorer operations may also be responsible for the loss of organisms (Montagna et al, ). However, Montagna et al () found that communities collected via boxcorer and multicorer at the same locations were very different, making comparisons between methods extremely difficult.…”

“…The washing and sieving of samples on the boat deck during boxcorer operations may also be responsible for the loss of organisms (Montagna et al, ). However, Montagna et al () found that communities collected via boxcorer and multicorer at the same locations were very different, making comparisons between methods extremely difficult. In spite of these large differences among sampling methods, much more variability was associated with seep communities compared to sampling methods, highlighting the large amount of variability associated with seeps (Figure ).…”

“…The top 15 cm of sediment was sieved onboard immediately after collection on 300‐µm mesh, and all material retained on the sieve was preserved with the addition of 10% buffered formalin. Although DGoMB samples included sediment 5 cm below DWH samples, it has been shown that little information on benthic community structure is obtained below 10 cm in sediment depth (Montagna, Baguley, Hsiang, & Reuscher, ). All specimens were sorted and identified to lowest taxonomic level by various taxonomic laboratories (Boland & Rowe, ; Washburn et al, ).…”

“…Total macrobenthic abundance was the only univariate parameter statistically analyzed for all samples using a one‐way nested ANOVA with habitat (seep, background) nested within depth zone (<1,000 m, 1,000–2,000 m, >2,000 m) as variables. Although cores collected within one multicore drop are considered pseudoreplicates, Montagna et al () found that more variance in macrobenthic abundance and richness was associated with cores within a deployment compared to cores among replicate deployments. Thus, cores within a multicore deployment were treated as replicates for statistical analyses.…”