INTRODUCTION

Marine benthic communities in the inshore coastal zone are subject to both natural and anthropogenic sources of disturbance. Shallow-water habitats in Long Island Sound (LIS) regularly experience periodic disruption of seafloor sediments that result from seasonal storm and wind events and from the action of tidal currents and waves. Aquaculture activities, such as shellfish harvesting, represent intermittent sources of anthropogenic seafloor disturbance. Hydraulic dredging uses pressurized seawater to loosen clams from surface sediments for collection by a dragged metal cage (Getchis, 2006; MacKenzie et al., 2002). Disruption of the seafloor, whether an outcome of natural or man-made processes, can physically displace surface sediments and dislodge resident organisms from the seabed, altering sediment characteristics and benthic community structure. Impacts on inshore habitat and recovery of benthic assemblages relate closely to the intensity, duration, spatial scale, and frequency of the disturbance (Dumbauld, Ruesink, and Rumrill, 2009).

Hydraulic shellfish dredging to harvest northern quahog or hard clam, Mercenaria mercenaria (Linnaeus 1758), has been conducted in Connecticut waters since the 1950s. Nearly 168 square miles (43,511 ha) of LIS seafloor is currently leased to harvesters for shellfish aquaculture (Getchis, 2005). Clams are collected from leased beds over a period of hours or days and effort varies with type of harvest activity, crop cycle, and size of cultivated beds (Dumbauld, Ruesink, and Rumrill, 2009). Postharvest, beds remain undisturbed for 2 to 5 years while small clams left behind by the dredge grow to market size. Reseeding of beds occurs naturally through seasonal settlement and recruitment or by transfer of clams from other locations. Commercial aquaculture produces more than 450,000 bushels of clams annually and contributes maritime jobs, revenue, and a source of sustainable, locally grown, high-quality seafood to Connecticut's economy (Connecticut Department of Agriculture, 2014). Shellfish harvest activities in LIS have been a factor shaping benthic community composition in the inshore coastal zone for more than half a century.

Successional processes related to natural disturbance of benthic communities have been well studied in LIS (e.g., McCall, 1977; Rhoads, McCall, and Yingst, 1978; Sanders, 1958). Abundance and distribution of organisms vary according to the stage of postdisturbance recovery, with species typically exhibiting either an opportunistic or equilibrium life-history strategy (McCall, 1977). Recolonization of benthic fauna in disturbed areas is shaped by a variety of synergistic factors including environmental conditions (e.g., sediment type, hydrodynamics), life-history characteristics (e.g., feeding mode, adult motility), population processes (e.g., life span, generation time), seasonal cycles (e.g., reproductive cycles, settlement), and biotic interactions (e.g., predation, competition) (Rosenberg, 2001; Zajac, Whitlatch, and Thrush, 1998; Zajac and Whitlach, 1982). The variable nature of the inshore environment contributes to the potential for more than one pattern of postdisturbance recovery (Probert, 1984). Long-term biological recovery, then, indicates re-establishment of an equilibrium community, rather than a wholesale return to the pre-existing assemblage, since even in the absence of disturbance, a community may not remain stable (Seiderer and Newell, 1999). Adaptive strategies displayed by infauna in response to seafloor disruption may account for spatial and temporal variability in population abundances (McCall, 1977).

Compositions of infaunal communities in LIS are closely tied to sediment texture and grain size (Zajac et al., 2000). Patterns of community structure and abundance are influenced by the relative proportions of silt and clay making up sediments, as deposit feeders show a preference for finer, organically rich sediments, whereas suspension feeders are generally more numerous in slightly coarser sediments (Sanders, 1956). A study of soft-sediment communities by Sanders (1956) identified 119 species of benthic fauna and a dominant “Nephtys incisa–Yoldia limulata” community type in LIS's central basin. Another study, conducted south of New Haven Harbor, identified three successional LIS communities: opportunists found in high numbers during initial recolonization, e.g., Streblospio benedicti, Capitella capitata, Ampelisca abdita; species at peak abundance during the mid-stage of succession, e.g., the bivalves Nucula proxima and Tellina agilis; and climax community species with low population densities, e.g., polychaete Nephthys incisa and bivalve Ensis directus (McCall, 1977; Rhoads, McCall, and Yingst, 1978). Observations from these studies suggest that although the species composition of LIS has remained generally consistent over time, changes in species abundance and dominance may occur (Zajac et al., 2000).

Recent studies in LIS have examined short-term effects of shellfish harvesting on benthic communities and sediment biogeochemistry by comparing newly harvested clam beds to sites dredged 1 to 2 years ago (Goldberg et al., 2012, 2014; Meseck et al., 2014). These studies, which showed little difference in species composition or abundance between dredged and adjacent uncultivated beds over weeks to months, suggest that dynamic inshore coastal areas favor establishment of communities resilient to disturbance. The present study examines longer-term ecological changes in benthic community structure on clam beds allowed to remain uncultivated over a broader time frame: 0 year (harvested just before sampling), 1 year postharvest, 2 years postharvest, and an inactive, not-dredged bed, undisturbed by harvesting for at least 10 years (10+).

METHODS

A field experiment was conducted on four leased shellfish beds where hard clams were periodically harvested. Study sites, distributed within a 60-ha area, reflect a range of dredging histories: 0 year (dredged just before sampling), 1 year postharvest, 2 years postharvest, and an inactive bed, not dredged for 10+ years. Sites were located within the same embayment, west of Charles Island, near Milford, Connecticut, within 1200 m of a midpoint 41°11.206′ N and −73°4.341′ W (Figure 1). Seawater depth in the study area varied with tidal cycle from 3.1 m to 4.9 m.

Figure 1. 

Study site in Long Island Sound (inset) off the coast of Milford, Connecticut, U.S.A. Four leased beds were sampled, each with different harvesting histories: 0 year, 1 year, 2 years, and 10+ years postharvest. Boxes show replicated plots for each treatment.

RESULTS

Environmental parameters were measured during each sampling event. Seawater temperatures ranged from 14.4 to 23.6°C over the study period. Mean salinity measured 25.8 ± 0.09 SE, and mean dissolved oxygen values were above saturation at 6.59 ± 0.37 SE parts per million.

PERMANOVA detected a significant effect of sediment grain size (p = 0.0001), dredging treatment (p = 0.0014), sampling week (p = 0.0001), plot (nested in treatment; p = 0.0001), and the interaction between treatment and week (p = 0.0001) on benthic assemblages. A pairwise comparison of treatments for individual weeks was performed to check this interaction, but the number of permutations was limited because of small sample size (24), and none of the pairwise comparisons was significant.

Mean grain size (Φ) measured 3.12 for 0 year, 2.76 for 1 year, 2.51 for 2 years, and 2.51 for 10+ years. Sites were characterized as fine (1, 2, 10+ years) to very fine (0-year) sand sediments. Sediment Φ size differed significantly among the dredging treatments (all p ≤ 0.013), except for the 10+- and 2-year site comparison (p = 0.41); therefore grain size was incorporated as a covariate into the PERMANOVA routine.

The ANOSIM test detected significant differences in benthic assemblages among all of the treatments (all p = 0.001), with the greatest difference between the 0-year and 10+-year sites (R = 0.872). Differences between the 1- and 2-year sites, although significant, were less pronounced, and had the lowest R value of all paired comparisons (R = 0.163). All other paired treatments had intermediate R values (0.381–0.544). These differences were apparent in both the MDS plot (not shown) and the CAP plot (Figure 2a). CAP analysis was used to identify which species were correlated with each dredging treatment. Species that increased or decreased in abundance in association with the 0-year or 10+-year sites are listed in Table 1. Presence of the bivalves M. mercenaria, Mulinia lateralis, Mya arenaria, Nucula spp., Yoldia limatula, gastropods, Haminoea solitaria, Illyanassa trivittata, Turbonilla spp. polychaetes Nepthys spp., and Pectinaria gouldii was correlated (>0.25) with the 0-year site. Organisms associated with the 10+-year treatment included the amphipods Ampelisca spp., Leptocheirus pinquis, bivalves Crassostrea virginica, E. directus, crustacean Xanthidae spp., echinoderm Leptosynapta spp., gastropod Crepidula fornicata, nemertean Lineus spp., and polychaetes Clymenella torguata, Drilonereis spp., Melinna cristata, and Phylodoce spp. Several species associated with the 1- and 2-year communities also overlapped with the 10+-year site, suggesting a transition in species composition with time after harvest. The gastropod Crepidula fornicata was correlated with both the 10+-year and 1-year sites. Species dominating the 2-year treatment included several species also associated with the 10+- year (H. solitaria, Mya arenaria, P. gouldii) and 1-year sites (Melinna cristata). The bivalve Anadara transverse and the isopod Edotea spp. were correlated with 2-year site only.

Figure 2. 

Canonical analysis of principal components identifying organisms associated with (a) dredging treatment and (b) sampling month.

Table 1. 

Dominant species associated with 0-year and 10+-year postharvest treatments, as determined by canonical analysis of principal coordinates (Spearman correlations >25%).

Sampling month influenced the composition of the benthic assemblages, as shown in the CAP plot (Figure 2b). ANOSIM results based on sampling month indicate that benthic assemblages for all months differed significantly (p ≤ 0.027) except for comparisons between August and September and September and October. Seasonal differences in benthic assemblages were most pronounced between the months of July and October (p = 0.001; R = 0.312).

PERMDISP found that dispersion in benthic communities within treatments across the sampling period was greatest in the 10+-year treatment, whereas the 0-year assemblages showed the lowest dispersion. The 1- and 2-year treatments showed an intermediate level of within-treatment variability.

Repeated-measures analysis detected significant differences between dredging treatments for number of species, number of individuals, Margalef richness, Shannon diversity, Pielou evenness, biovolume, mollucan phyla, and also for TOM (Table 2). The number of individuals at the 10+-year site was significantly lower than at the other treatments (Figure 3a). The 1- and 2-year sites contained significantly more species than the 0-year and 10+-year sites (Figure 3b). Significantly more mollusks were found on the 0-year bed vs. the other locations, and the lowest number of mollusks was observed at the 10+-year site (Figure 3c). Margalef indicated significantly lower species richness at the 0-year site compared with the 10+-, 2-, and 1-year sites, which did not differ. The Shannon and Pielou indices indicated lower diversity and evenness at the 0-year compared with the 10+-year site. Biovolume of organisms was significantly higher at the 0-year site vs. the 10+-, 2-, and 1-year sites, whereas biovolumes at the 1- and 2-year sites were significantly greater than at the 10+-year location. TOM (mg/g) was significantly greater at the 0-year site as compared with the 1-, 2-, and 10+- year sites (Figure 4), whereas TOM concentrations were significantly lower at the 10+-year site compared with the 0-, 1-, and 2-year sites. TOM at the 1- and 2-year postharvest sites did not differ.

Table 2. 

Results of repeated-measures tests for ecological indices and total organic matter comparing clam beds with different dredging histories. Number indicates which year was consistently greater in the pairwise comparison; p-values indicated by * ≤ 0.05, ** ≤ 0.01, *** ≤ 0.001, ns = not significant.

Figure 3. 

Mean number of individuals (a), mean number of species (b), and mean number of mollusks (c) on clam beds with different dredging histories from June to October 2011.

Figure 4. 

Mean total organic matter (mg/g) for clam beds with different dredging histories from June to October 2011.

DISCUSSION

This study observed significant differences in benthic community structure related to the hydraulic dredge harvesting history of clam beds. Some differences in benthic assemblages were apparent immediately after dredging disturbance (e.g., the newly harvested 0-year site), whereas other differences were related to the time interval since dredging. The recently harvested 0-year site was very similar to the predominant central basin, LIS benthic community described by Sanders (1956) over 50 years ago, typically composed of suspension (P. gouldii [Cistenoides old name], Mulinia lateralis) and deposit feeders (Nephtys incisa, Nucula proxima, Y. limatula).

Natural or anthropogenic changes in seafloor conditions and sediment characteristics after disturbance may reduce habitat suitability for certain organisms while offering a more favorable environment for others (Sparks-McConkey and Watling, 2001). Recently disturbed environments are typically recolonized by pioneer species with rapid reproductive cycles, brief life spans, and high dispersal rates (Rhoads, McCall, and Yingst, 1978; Thistle, 1981). Periodic disruption of the seafloor is necessary for some species to persist in marine soft-bottom communities (Thistle, 1981). In a South Carolina study, organisms were more abundant in mechanically harvested tidal creeks vs. unharvested reference creeks within a few weeks of harvest (Maier et al., 1998). In LIS studies examining short-term effects of hydraulic shellfish dredging, few significant differences were observed in direct comparisons between adjacent dredged and not-dredged clam beds in the first few weeks and months immediately after harvest (Goldberg et al., 2012, 2014). The present study observed higher biomass and quantity of individual organisms during the first few years after harvest, followed by a reduction in biovolume and overall abundance within 10+ years postdredging.

Bivalve mollusks comprised the largest proportion of species associated with the 0-year site, and were significantly more abundant on this newly dredged bottom. Suspension-feeding clams are considered early colonizers in LIS, known to settle in nearshore, high-current, disturbance-prone areas (Rhoads, McCall, and Yingst, 1978). A previous study on a Connecticut shellfish lease found greater numbers of newly settled hard clams on recently dredged seabed compared with nearby not-dredged bottom (Goldberg et al., 2014). Increased numbers of settling bivalves have been observed after harvest at other locations along the eastern seaboard (e.g., Kyte, Averill, and Hendershott, 1976; Pfitzenmeyer, 1972a,b; Rice, Hickox, and Zehra, 1989). A pulse of bivalve settlement was observed, primarily Nucula spp., at the newly harvested site during mid-July, 3 weeks after dredging, which may indicate a localized settling event related to recent bottom cultivation. A spike in young-of-the-year Y. limatula was similarly observed during late July. In a study of sediment texture and molluscan abundance in LIS, Franz (1976) observed low evenness and high dominance in the Nucula/Nepthys/Yoldia infaunal community. The presence of very fine sediments (3.12 Φ) at the 0-year site may account for the high abundance of newly settled bivalves observed in this study, as both Nucula spp. and Y. limatula are highly associated with very fine sand bottom (Franz, 1976; McCall, 1977; Sanders 1956).

Cultivation of sediments by hydraulic dredging may alter the physical and chemical properties of bottom sediments in such a way that settlement, recruitment, or survival of young shellfish is enhanced. Slightly smaller sediments at the 0-year site, compared with the other locations, may reflect a geographic gradient of decreasing particle size, or may indicate resuspension of sediments into the water column by recent dredging activity and subsequent size grading as material resettled on the seafloor (Sparsis, DeAlteris, and Rice, 1993). Tilling of the seafloor by harvesting may resuspend, disperse, and redistribute finer sediment components (Maier et al., 1998), potentially increasing TOM and food availability for bivalves. Higher TOM concentrations at the 0-year site vs. the other locations may be attributed to a variety of factors including (1) generally higher deposition of TOM at the 0-year site, (2) remixing of not-yet-decomposed organic matter back into sediments during dredging, (3) less bioturbation activity resulting in a slower rate of degradation, or (4) a combination of these factors. Conversely, greater biodiversity at the 10+-year site suggests more bioturbation, greater reworking, and a higher rate of degradation of organic matter in the sediments, which would account for consistently lower TOM than found at the other sites. In a study in Italy, Fiordelmondo et al. (2003) observed increased nutritional availability of sediment organic matter after clam harvesting. Deposit-feeding mollusks, such as Nucula spp. and Y. limatula, obtain nourishment from processing organically rich surface sediments (Sanders, 1956). Higher numbers of newly settled mollusks on recently harvested bottom may reflect a habitat preference for slightly finer sediments, increased survival due to the presence of greater food availability via elevated TOM, or reflect short-term changes to sediment chemistry, texture, or particle size mediated by harvesting activities.

The 1- and 2-year postharvest sites contained a significantly greater number of species than found at the newly harvested 0-year and fallow 10+-year locations, suggesting ongoing recolonization during the first few years postharvest. When disturbance occurs at regular intervals, recruitment and productivity may remain high (Rhoads, McCall, and Yingst, 1978). Although recolonization of disturbed bottom generally occurs rapidly in the inshore coastal zone (Sparks-McConkey and Watling, 2001), recovery of benthic communities can continue for months (Thistle, 1981). In a study of hydraulic clam dredging in Canada, an increase in abundance of nontarget species, a decline in taxonomic distinctness, and continued recolonization of the seafloor was observed up to 2 years postharvest (Gilkinson et al., 2005). Species richness, diversity, and evenness indices did not differ among the 1-, 2-, and 10+-year beds, indicating a greater similarity between these communities, vs. the recently harvested bed. Our study suggests that 1 to 2 years after harvest represents a transition period between colonization by early pioneer communities and establishment of longer-term equilibrium assemblages, when species representing both life-history strategies may co-occur.

The community of species correlated with the 10+-year site reflected a broader range of phyla and functional ecological roles than those associated with the 0-year site. Ecological indices (species richness, diversity, evenness) were higher at the fallow 10+-year site than the newly harvested location. Environments that remain uncultivated for an extended period would be expected to surpass initial stages of recolonization and include species that require a longer time frame to establish and more constant conditions to be sustained. When harvesting occurs intermittently, a wider variety of organisms may accumulate over time, reflective of natural variations in annual year class strength and seasonal abundance cycles. Other LIS studies of successional processes have also noted reduced diversity on newly recolonized seafloor and more varied and longer-term communities in less frequently disturbed locations (McCall, 1977). Gradual migration of adults into disturbed patches may also contribute to more complex community composition over time (Zajac and Whitlach, 1991). The assemblage of species found at the 10+-year site indicates progression toward establishment of an equilibrium community. Molluscan species like Crassostrea virginica and Crepidula fornicata, which attach and adhere to substrate, were closely associated with the fallow 10+-year site, where the bottom remained more stable. Polychaetes represented a greater proportion of the most common species at the 10+-year site, suggesting a proliferation of these species in undisturbed sediments. Older, established communities typically support a greater diversity of organisms, but at a lower carrying capacity than opportunistic pioneer communities. However, the presence of regular disturbance events in LIS habitats suggests that few communities consistently achieve or sustain mature successional composition (Rhoads, McCall, and Yingst, 1978).

Benthic community composition was significantly influenced by month of sampling, independent of dredging history. Species assemblages showed distinct patterns associated with sampling month that were apparent regardless of harvest year. Rising spring and summer seawater temperatures trigger spawning events, larval settlement, and recruitment that increase benthic population size, whereas declining autumn temperatures signal a slowdown in overall productivity and abundance. Previous short-term studies of early summer hydraulic dredging in LIS found sampling month to have a greater influence on seafloor community composition than one-time harvesting when followed over a single growing season (Goldberg et al., 2012, 2014). Natural shifts in dominant taxa and benthic community structure related to season are generally apparent regardless of the stage of postharvest recovery (Maier et al., 1998).

CONCLUSIONS

Successional changes in benthic community composition related to hydraulic harvesting history were observed on LIS clam beds. The present study suggests that biomass of organisms on clam beds is greatest in the first few years after harvest. Greater numbers of newly settled bivalves on recently dredged bottom may be related to the higher levels of TOM observed there. Benthic assemblages on sites undisturbed for 1 and 2 years postharvest contained the highest number of species and represented a transition in composition between the 0-year and 10+-year postharvest communities. Recently dredged bottom (0 to 2 years) contained more individuals and higher biomass than the fallow (10+ years) location. Left uncultivated for 10+ years, benthic assemblages will revert gradually through succession to communities with fewer individuals per species but generally greater species diversity and evenness as compared to newly harvested bottom. Cultivation of sediments may bring about short-term changes to the consistency of sediments that influence species composition. Frequent natural disturbance and intermittent dredge harvesting on LIS clam beds in the inshore coastal zone may result in a patchwork of ecological communities with differing composition and at varying stages of succession.