The importance of fine-scale flow processes and food availability in the maintenance of soft-sediment communities

Although the association between soft-sediment invertebrates and a specific sediment type hasbeen documented for many habitats, most studies have been correlative and have failed to convincinglydemonstrate any single mechanism to explain this association. Sediment type has generally beencharacterized by grain size, however, many other potential causal factors correlate with grain size,including organic content, microbial content, stability, food supply, and larval supply. One hypothesis foranimal-sediment associations is that settling larvae are transported as passive particles and are sorted intodifferent sedimentary habitats much like sediment grains.To test the hypothesis that near-bed hydrodynamics may modify larval settlement, field and flumeexperiments were conducted where larval settlement was compared between microdepositionalenvironments (small depressions) and non-trapping environments (flush treatments). Depressions havebeen observed to trap passive particles, and these experiments were therefore designed to test whethersettling larvae would be trapped in depressions like passive particles. Flume flow simulations were carriedout with the polychaete Capitella sp. I and the bivalve Mulinia latera/is. Experiments with flush anddepression sediment treatments were conducted in the absence of the potentially confounding effects ofsuspended sediment and organic matter and therefore offered a highly controlled, explicit test of passivehydrodynamic deposition of larvae in depressions. Although larvae of both species were generally ableto actively select a high-organic sediment over a low-organic alternative with a comparable grain size,elevated densities of both species were observed in depressions for a given sediment treatment. Thus,both species appeared to be vulnerable to hydrodynamic trapping. M. latera/is larvae, however, oftenmade a "poor choice" by settling in high numbers in depressions containing the low-organic sedimentwhile Capitella sp. I larvae were generally able to "escape" from depressions if the sediment wasunsuitable. In field experiments carried out at Station R in Buzzards Bay, Massachusetts, significantlyhigher densities of Mediomastus ambiseta juveniles, spionid polychaete juveniles, bivalves, gastropodlarvae, and nemerteans were observed in depressions compared with flush treatments over 5 relativelyshort experimental periods (3 or 4 days each) during the summer of 1990. Of the abundant taxa, onlyCapitella spp. was not significantly more abundant in depressions compared with flush treatments,although numbers tended to be higher in depressions. Experiments were conducted over a short timeperiod to minimize potential biological interactions between taxa and reduce the likelihood that organicmaterial would accumulate in depressions and provide a cue for settling larvae. Thus, higher numbersin depressions suggest that larvae were passively entrained. These flume and field experiments suggestthat near-bed hydrodynamics may modify settlement at some scales, and that both active and passiveprocesses may operate in determining larval distributions in shallow-water, muddy habitats.In deep-sea ecosystems, the role of near-bed hydrodynamics is also of interest because of thepotential role that larval settlement in organic patches may play in maintaining the immense speciesdiversity characteristic of many deep-sea ecosystems. To try to understand the role of organic patchesin deep-sea communities, several investigators have used colonization trays containing sediments that havebeen treated in different ways. These experiments have been criticized in the past because the sedimentsurface in the trays was elevated above the bottom and may therefore have interfered with natural boundary layer flow. Flume simulations of flow over these colonization trays revealed serious flowartifacts generated by the trays, and that flow across the sediment surface of the trays was characterizedby turbulent eddies, accelerated velocities and boundary layer thickening. These sorts of flowcharacteristics would not be expected over natural sediments, and an alternative colonization tray wasdesigned to eliminate these artifacts.To test the hypothesis that different types of food patches would result in different types of larvalresponse, and determine how near-bed hydrodynamics may influence larval settlement, flush colonizationtrays filled with prefrozen sediment were deployed in tandem with artificial depressions south of St. Croix,U.S.V.I at 900 m depth. Colonization trays and artificial depressions were either unenriched or enrichedwith Thalassiosira sp. and Sargassum sp. two types of algae chosen to mimic natural food patches on thesea floor. Unexpectedly high densities of organisms colonized trays after only 23 days. The Thalassiosiratrays were colonized by high densities of a relatively low diversity, opportunistic fauna, Sargassum trayswere colonized by lower densities of a higher diversity fauna, and unenriched trays were colonized byvery low numbers of a very diverse fauna. All tray faunas were markedly different in composition fromthe natural, ambient fauna. These fmdings suggest that different patch types did, indeed, result in aspecialized faunal response to each of the "patch" types. Depressions on the sea floor provide a naturalmechanism for food patch formation because passive particles such as detritus and algae tend to beentrained in the depressions. To determine whether dominant colonizers would be entrained in depressionslike passive particles or could differentiate between depression "patch" types in a flow environment thatmight be expected to make active selection more difficult, artificial depressions were unenriched orenriched with Sargassum sp. or Thalassiosira sp. Total densities of organisms and densities of the mostabundant species were substantially lower in artificial depressions than in trays. Densities in Thalassiosiradepressions were lower than in Sargassum depressions and densities in unenriched depressions wereextremely low, suggesting that dominant colonizers were not passively entrained in depressions and thatcolonization was specialized and highly active for these taxa. A different fauna was also observed innatural depressions compared with flush sediments, suggesting that natural depressions do contribute tospecies coexistence. Long-term tray deployments designed to test whether different faunas would bepresent in "patches" of different ages indicated that time may also play an important part in a deep-seapatch mosaic.

This was funded by NSF and ONR, NOAA, NSERC (Canada), WHOIOcean Ventures Fund and the WHOI Ditty Bag Fund.

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy at the Massachusetts Institute of Technology and the Woods Hole Oceanographic Institution January 1993