Comparison of two benthic assemblage sampling gears for use on intertidal oyster reefs in Louisiana

Introduction

Loss of estuarine biodiversity due to climate change and anthropogenic activities may negatively impact estuarine habitats’ provisioning of ecosystem functions and services (Lotze et al., 2006). To address this loss, many local and global initiatives focus on preserving, restoring, and enhancing ecosystems to maintain biodiversity in estuaries (United Nations Environment Programme (UNEP), 2021). The identification and development of robust monitoring tools and metrics enable tracking and understanding of the impacts of these management efforts on biodiversity (Flannery & Przeslawski, 2015).

Monitoring biodiversity in estuaries is challenging due to the diverse and often structurally complex habitats (e.g., shellfish reefs) impacting sampling gear efficiency (Baker & Minello, 2011; Flannery & Przeslawski, 2015; Mihoub et al., 2017; La Peyre et al., 2021). The use of different gears to sample estuarine benthic assemblages (hereafter assemblages), such as ponar grabs, trays, or suction samplers can result in different values for taxa abundances, assemblage structure, diversity, and richness (Stoner et al., 1983; Slack, Ferreira & Averett, 1986; Keklikoglou et al., 2019). One of the more complex habitats within estuaries are reefs built by the Eastern oyster Crassostrea virginica. These reefs support significant biodiversity but estimates of assemblage abundance and richness vary between studies that have sampled reefs of different complexity and/or used different gears (Wells, 1961; Coen & Grizzle, 2007; La Peyre et al., 2019). For example, a meta-analysis of studies from coastal Texas through Florida documented 115 fish and 41 decapod crustacean species with densities as high as 2,800 m⁻² and as many as 52 species per reef, but the results differed significantly based on the gear types used (La Peyre et al., 2019). These studies highlight that gear comparison is complicated by habitat characteristics which can affect a gear’s ability to capture a representative proportion of the target assemblages (efficiency) and specific taxa and size classes (selectivity) across habitats (Wells, 1961; Stoner et al., 1983; Slack, Ferreira & Averett, 1986; Coen & Grizzle, 2007; Keklikoglou et al., 2019; La Peyre et al., 2019). Consequently, researchers and managers still search for gear that maximize the taxa richness captured, minimize sampling time and effort, and promote more consistent biodiversity assessment across habitats, locations, and seasons (Flannery & Przeslawski, 2015).

To date, there are no studies in oyster reefs comparing the selectivity and efficiency of trays and suction sampler gears designed to sample benthic assemblages. Trays are often used for sampling biodiversity on oyster reefs (La Peyre et al., 2019). They create minimal reef disturbance, minimize loss of escaping organisms, and allow the collection of organisms that reside in reef interstitial spaces (Beck & La Peyre, 2015). However, trays are time-consuming to deploy, have a high risk of gear loss due to long deployment times, and potentially bias results due to the added structure of the tray (Beck & La Peyre, 2015). In contrast, suction samplers require less time and have a lower chance of gear loss because samples are collected in one field event with no deployment time required. Suction samplers are often limited to shallow water habitats and capturing size classes as large as the suction diameter. Additionally, suction samplers create a noise disturbance from the motor, which can impact biodiversity sampling by scaring away or altering behavior of noise-sensitive species (Flannery & Przeslawski, 2015). It is also unclear how efficiently suction samplers capture organisms that reside in the interstitial space within reef substrate. Until recently, the suction sampler was used primarily on soft bottom (True, Reys & Delauze, 1968) but is gaining popularity for sampling oyster reefs (Pinnell et al., 2021; Pollack, Palmer & Williams, 2021).

In this study, we compared two sampling gears, trays and suction samplers, in collecting benthic biodiversity on natural intertidal C. virginica oyster reefs. By comparing four common metrics of biodiversity calculated from each sampling gear across three reefs, we hypothesized that tray samples would consistently yield higher taxa abundance, richness, Pielou’s evenness, and Shannon-Wiener diversity than suction samplers. This study provides insight into potential biases or limitations of each gear, leading to more accurate and comparable data for future studies. Comparing benthic biodiversity sampling gears informs management and conservation strategies for oyster reefs, which hold significant ecological and economic value.

Materials and Methods

Study site

The study area encompassed approximately 1 km² of coastal habitats near Cocodrie, Louisiana which included Spartina alterniflora dominated marsh, microtidal channels, bayous, and ponds interspersed with oyster reefs (Fig. 1A). Over the past decade (2012–2022) water depth ranged from 1.0–2.5 m (1.7 ± 0.0009, n = 52,497), water temperature ranged from 1.5–35.2 °C (23.0 ± 0.01, n = 309,313), and salinity ranged from 0.6–25.1 PSU (9.5 ± 0.01, n = 241,430; http://weatherstations.lumcon.edu/index.html).

Results

Throughout collection days water depth (1.85 ± 0.01 m, n = 403), temperature (29.15 ± 0.04 °C, n = 384), salinity (5.96 ± 0.06 PSU, n = 384), and dissolved oxygen (5.42 ± 0.05 mg L⁻¹, n = 384) were typical for this region during the summer (Table S2).

Across the reefs, reef substrate (live + dead oyster material) volume ranged from 4–48 L m⁻², live oyster densities ranged from 0–144 ind m⁻², and cluster volume ranged from 3.2–32 L m⁻² (Table S3). Reef 1 had the lowest live oyster density (0 ± 0 ind m⁻²) and substrate volume (18.29 ± 3.36 L m⁻²) while reef 2 had the highest live oyster density (73.14 ± 21.20 ind m⁻²) and substrate volume (28.57 ± 5.27 L m⁻²).

Trays consistently contained significantly higher abundances compared to suction samples (Kruskal Wallis tests: Reef 1 χ² = 8.34, df = 1, p = 0.004; Reef 2 χ² = 8.31, df = 1, p = 0.004; Reef 3 χ² = 8.34, df = 1, p = 0.004) and no significant differences in abundance were detected among reefs for either gear (Fig. 2A). Similarly, trays consistently contained significantly higher richness than suction samples (Kruskal Wallis tests: Reef 1 χ² = 8.49, df = 1, p = 0.004; Reef 2 χ² = 8.49, df = 1, p = 0.004; Reef 3 χ² = 6.70, df = 1, p = 0.0096; Fig. 2B). In trays, there was significantly lower richness at reef 1 compared to reef 2, and no significant differences between these reefs and reef 3 (Post Hoc Dunn tests: Reefs 1–2 Z =−2.75, p adj. = 0.02). No significant differences in richness among reefs were detected from suction samples. Trays consistently contained significantly lower Pielou’s evenness compared to suction samples across reefs (Kruskal Wallis tests: Reef 1 χ² = 7.52, df = 1, p = 0.006; Reef 2 χ² = 8.37, df = 1, p = 0.004; Site 3: χ² = 8.34, df = 1, p = 0.004; Fig. 2C). In suction samples, there was significantly higher evenness at reef 1 compared to reef 2, and no significant differences between these reefs and reef 3 (Post Hoc Dunn tests: Reefs 1–2 Z = 2.58, p adj. = 0.02). No significant differences in evenness among reefs were detected from tray samples. For tray and suction samples, Shannon-Wiener diversity did not differ significantly between gears at reefs 1 and 2 but was significantly higher in suction samples than trays at reef 3 (Kruskal Wallis test: χ² = 8.43, df = 1, p = 0.004; Fig. 2D). However, no significant differences in Shannon-Wiener diversity were detected among reefs for either gear.

Discussion

The collection of comparable data across projects and locations enables the development of robust databases to inform biodiversity protection (Flannery & Przeslawski, 2015). Here, we compared suction samples and trays to assess benthic biodiversity metrics on and between oyster reefs. We found significant differences in biodiversity metrics between gears and across reefs that varied by metric. Gear selectivity and efficiency, which can differ by habitat characteristics, likely contribute to differences in gear outcomes, further indicating that tray and suction sampling are not comparable biodiversity sampling gears on oyster reefs. Using two gear types to compare selected biodiversity metrics across a complex estuarine habitat, we highlight the importance of understanding and acknowledging how sample gear efficiency, and selectivity may impact biodiversity estimates and comparisons. These findings may be beneficial to managers developing monitoring and adaptive management plans and help inform conservation decisions by acknowledging uncertainty or variation in biodiversity estimates resulting from sampling protocols and metric calculations.

Biodiversity metrics often differ based on the targeted assemblages collected. Here, the average organism abundance of 1,806 ind m⁻² (SE ± 318 m⁻² n = 18) and 23 total taxa collected in trays was higher than majority of average fish and/or decapod crustacean abundance (ranging from 57–1,579 ind. m⁻²) and total taxa (ranging from 8–22) per study that used trays as sampling gear in an oyster reef assemblage meta-synthesis across the U.S. Gulf Coast (USGC; La Peyre et al., 2019). Only two studies in the meta-synthesis, both from Texas reefs, reported higher numbers of abundance (2,856 ind m⁻²; Rezek et al., 2017) or total taxa (25; Blomberg et al., 2018). The meta-synthesis studies only included fish and/or decapod crustaceans, while here we also include smaller taxa such as amphipods, isopods, and polychaetes. When removing the smaller taxa from this dataset, average abundance (698 ± 51 m⁻² n = 18) and total taxa (15) were more comparable to studies in La Peyre et al. (2019). Additionally, tray modifications, such as lining trays with mesh drawstring bags, can impact gear efficiency. We lined trays with a smaller mesh size (1 mm) compared to past studies which likely contributes to the higher abundance observed (Beck & La Peyre, 2015). The inclusion of amphipods, which were numerically dominant in our trays along with our gear modification, likely drove the higher abundances and richness captured but does not explain lower suction sampling results in this study as other studies have largely focused on smaller fauna when suction sampling.

The use of suction sampling in reef environments in other studies is sparse but the assemblage metrics are similar to trays in the USGC likely because of the epifauna and infauna target assemblages (La Peyre et al., 2019; Pinnell et al., 2021; Pollack, Palmer & Williams, 2021). In contrast, our suction sample abundances ranged from 17–204 ind m⁻² and richness from 2–10 which are lower than the few similar studies’ ranges of abundance (241– 8,800 ind m⁻²) and richness (6–12) per suction sample (Pinnell et al., 2021; Pollack, Palmer & Williams, 2021). Differences in gear efficiency can create discrepancies across gears and is further impacted by gear modifications, such as modifying the amount of time used to suction a given area (Brown, Schram & Brussock, 1987). We suctioned 0.23 m⁻² area for 15 s, which is lower than 0.0625 m⁻² area for 30 s in Pinnell et al. (2021), likely contributing to the lower abundance and richness in this study. Although Pinnell et al. (2021) collected organisms in a different geographic region (Pacific coast) compared to this study (USGC), the cutoff size (>500 μm vs >1 mm in this study) of organisms and gear modifications used likely drove the comparatively higher assemblage metrics. Additionally, Pollack, Palmer & Williams (2021) combined gears on Texas reefs, using suction samplers to suction reef habitat placed within trays, whereas we used suction samplers and trays separately. Differences in reported abundance and richness between studies may partially stem from variations in gear design and methods (gear combinations, mesh size), resulting in differences in gear selectivity.

Although there were many taxa caught by both sampling gears across reefs, suction samplers and trays captured some taxa that the other did not (tray = 8, suction = 6). Generally, suction samples captured more sessile invertebrates (e.g., Ameritella mitchelli) associated with mud-bottom habitat than trays, while trays captured more mobile reef-associated taxa that suction samples did not (e.g., Gobiesox strumosus, Table 1) across reefs. The average number of taxa captured only in trays (mean _richness = 3.77 ± 0.68 m⁻², n = 144 _{individuals per sample}) were more abundant than average taxa captured only in suction samples (mean_richness = 0.56 ± 0.17 m⁻², n = 108_{individuals per sample}). These gear-specific taxa contributed to the higher tray abundance and richness compared to suction samples and is logical based on how each gear operates. Our two-week tray deployment time was not long enough for sessile invertebrates to settle and grow large enough to identify. Additionally, mobile taxa can escape the enclosed suction sampling area on uneven oyster reef substrate, likely explaining the fewer mobile taxa caught. In suction sampling studies within seagrass, a drop net is used to prevent mobile taxa from escaping and is not inherently a limitation of suction sampling (Ralph et al., 2013). However, it appears the use of drop nets with suction sampling has not translated to reefs yet. Reporting gear modifications and standardized (m⁻²) biodiversity metrics allows for the comparison of assemblages across reefs and gears for biodiversity monitoring.

Our findings indicate tray and suction sampling are not comparable within a single oyster reef nor across oyster reefs. One explanation for the differences across oyster reefs within a single gear may be due to differences in reef characteristics (e.g., structural complexity), which can drive differences in the assemblage patterns among reefs (Pinnell et al., 2021). However, tray and suction samples detected contrasting differences among reefs for richness and evenness and different relative rankings across space (e.g., tray average richness: reef 2 > reef 3 > reef 1, suction average richness: reef 3 > reef 2 > reef 1; Table S3). Additionally, when considering unique taxa (i.e., a taxon represented by a single individual within a sample), on average 51% of the taxa collected in a suction sample were unique, whereas only 17% of taxa within a tray were unique. More unique taxa collected in suction samples was likely due to a lack of abundance of common taxa also collected in trays in higher abundance. All trays had at least one numerically dominant taxa whose abundance was more than two times the sample-specific mean, but this only occurred in 44% of suction samples. The lower proportion of unique taxa and higher proportion of numerically dominant taxa in trays here explain the differences in Pielou’s evenness and Shannon-Wiener diversity observed between the gears. Contrasting gear results of richness and abundance with Pielou’s evenness led to the similarity in Shannon-Wiener diversity between gears. The samples collected via trays consistently contained more individuals, more dominant taxa, and had significantly different patterns among reefs compared to suction samples, highlighting the incomparable biodiversity metric results generated from these gears on oyster habitat.

Conclusions

Sampling gear choice depends on factors such as time frame, equipment, funding, field standards, and target assemblages. Our study demonstrated that oyster reef-associated biodiversity metrics varied significantly between gears and across reefs. Trays captured more individuals and taxa overall, including dominant taxa, while suction samplers collected more unique taxa. While Pollack, Palmer & Williams (2021) captured taxa abundance and richness using suction samples that aligned with previous data captured using trays, we did not. The discrepancy highlights the potential influence of other factors, such as habitat complexity, area sampled, and suction duration, on estimates of biodiversity generated from suction and tray samples. Comparing and standardizing field methods, rather than continually developing new sampling approaches, may be more beneficial for reconciling discrepancies to future research and monitoring. Moreover, further research on the impacts of gear modifications, environmental factors, and habitat characteristics may enhance sampling gear decision making. Understanding biodiversity metrics and associated trade-offs for selecting a gear remains critical because the gear’s effectiveness is dependent on the project goals and target assemblages (Yi et al., 2012). Our study contributes to informed decision-making by managers, researchers, and practitioners about gear selection and data interpretation for targeted studies and monitoring programs.

Supplemental Information