Mud Erosion Research Articles

Morphodynamics of the Wadden Sea and its barrier island system

The Wadden Sea and its associated barrier island system exhibit highly dynamic behaviour. Of major concern is the movement of water and air and the transport, erosion and deposition of sand and mud. These processes result in an ever-changing morphology (topography/bathymetry) of the islands, tidal channels, inter-tidal shoals and tidal flats. This dynamic development of the shape and nature of the Wadden area forms together with the biotic systems, the present Wadden system. The morphodynamic development of the Wadden Sea is influenced by changing environmental conditions e.g. sea-level rise as well as by human interferences. For the management and protection of the Wadden system knowledge on the morphodynamic development is essential. However, our present knowledge is not sufficient to predict the effects of human interferences under different climate change scenarios in sufficient detail and accuracy. This paper identifies the existing knowledge gaps, based on a review of the state of the art on morphodynamics of the Wadden Sea that is confronted with major requirements from a coastal zone management point of view. The identified knowledge gaps have to inspire and stimulate research in the fields of the large-scale sediment budgets, morphodynamic changes at smaller scales, processes and mechanisms of sediment transport, erosion and deposition and modelling tools.

Dynamic Mud Deposition In A Tidal Environment: The Record of Fluid-Mud Deposition In the Cretaceous Bluesky Formation, Alberta, Canada

Abstract Mudstone layers in the tide-dominated Bluesky Formation (i.e., the “mud drapes”) have enormously variable sedimentary characteristics: they range from 0.1 to 20 cm thick, can be homogeneous or internally stratified, and can have sharp or gradational upper and lower contacts. Based on recent flume studies, this diversity is interpreted to reflect the wide range of suspended-sediment concentrations (SSC; 100 g L−1) and flow velocities (slack water to several meters per second) that are typical in tidal environments. Four recurring mudstone types are present, each of which is interpreted to have formed under distinct depositional conditions: (1) unstratified mudstone Type 1 (UM1): “classic” mud drapes generated by slow settling from suspensions with SSC values less than 1 g L−1; (2) stratified mudstone Type 1 (SM1): cross-stratified mudstone layers deposited by turbulent or transitional turbulent flows with appreciable current speeds (> 0.2 ms−1) and moderate SSCs (1–10 g L−1); (3) stratified mudstone Type 2 (SM2): horizontally laminated mudstone layers formed under conditions of transitional plug flow with moderate to high SSCs (1–100 g L−1) and appreciable currents (> 0.2 ms−1); and (4) unstratified mudstone Type 2 (UM2): thick mudstone laminae and beds (> 2 mm thick), with no internal lamination and common soft-sediment deformation, deposited by unstable plug flow or quasi-laminar plug flow, with moderate to high SSCs (1–1000 g L−1) and current speeds ranging from slack water to 1–2 ms−1. UM2 and SM2 mudstone types are equivalent to what recent studies have interpreted as the lithified product of fluid mud and may have been deposited dynamically, under current speeds above the threshold of mud erosion in clear-water flows. This indicates that, in high-SSC settings, “mud drapes” could form over significant portions of the tidal cycle (i.e., not only at slack water) and perhaps continuously over multiple tidal cycles if fluid-mud layers persist. Thus, tidal rhythmites might be poorly developed in areas with high SSC values.

Bottom friction and erosion beneath long-crested and short-crested nonlinear random waves

A practical approach for calculating the bottom shear stress beneath long-crested (2D) and short-crested (3D) nonlinear random waves is provided. The approach is based on assuming the waves to be a stationary narrow-band random process and by adopting the Forristall (2000) wave crest height distribution representing both 2D and 3D nonlinear random waves. Results are presented for laminar, smooth turbulent and rough turbulent flow. Examples are also included to illustrate the applicability of the results for practical purposes using data typical for field conditions; the mobile layer thickness in sheet flow representing rough turbulent flow; erosion of mud representing smooth turbulent flow and deposition of mud representing laminar flow.

Stratigraphy and depositional setting of slurry and contained (reflected) beds in the Marnoso-arenacea Formation (Langhian-Serravallian) Northern Apennines, Italy

This work presents the stratigraphy and facies analysis of an interval of about 2500 m in the Langhian and Serravallian stratigraphic succession of the foredeep turbidites of the Marnoso-arenacea Formation. A high-resolution stratigraphic analysis was performed by measuring seven stratigraphic logs between the Sillaro and Marecchia lines (60 km apart) for a total thickness of about 6700 m. The data suggest that the stratigraphy and depositional setting of the studied interval was influenced by syndepositional structural deformations. The studied stratigraphic succession has been subdivided into five informal stratigraphic units on the basis of how structurally controlled topographic highs and depocentres, a consequence of thrust propagation, change over time. These physiographic changes of the foredeep basin have also been reconstructed through the progressive appearance and disappearance of thrust-related mass-transport complexes and of five bed types interpreted as being related to structurally controlled basin morphology. Apart from Bouma-like Type-4 beds, Type-1 tripartite beds, characterized by an internal slurry unit, tend to increase especially in structurally controlled stratigraphic units where intrabasinal topographic highs and depocentres with slope changes favour both mud erosion and decelerations. Type-2 beds, with an internal slump-type chaotic unit, characterize the basal boundary of structurally controlled stratigraphic units and are interpreted as indicating tectonic uplift. Type-3 beds are contained-reflected beds that indicate different degrees of basin confinement, while Type-5 are thin and fine-grained beds deposited by dilute reflected turbulent flows able to rise up the topographic highs. The vertical and lateral distribution of these beds has been used to understand the synsedimentary structural control of the studied stratigraphic succession, represented in the Marnoso-arenacea Formation by subtle topographic highs and depocentres created by thrust-propagation folds and emplacements of large mass-transport complexes.

Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea)

Multi-sensor tripod measurements in the high-turbidity area of the Belgian nearshore zone (southern North Sea) allowed investigating storm effects on near bed suspended particulate matter (SPM) concentrations. The data have shown that during or after a storm the SPM concentration increases significantly and that high concentrated mud suspensions (HCMS) are formed. Under these conditions, about 3 times more mass of SPM was observed in the water column, as compared to calm weather conditions. The following different sources of fine-grained sediments, influencing the SPM concentration signal, have been investigated: wind direction and the advection of water masses; the previous history and occurrence of fluffy layers; freshly deposited mud near the disposal grounds of dredged material, navigation channels and adjacent areas; and the erosion of medium-consolidated mud of Holocene age. Based on erosion behaviour measurements of in-situ samples, the critical erosion shear stresses have been estimated for different cohesive sediment samples outcropping in the study area. The results have shown that most of the mud deposits cannot be eroded by tidal currents alone, but higher shear stresses, as induced by storms with high waves, are needed. Erosion can however occur during storms with high waves. Data suggest that in order to obtain very high SPM concentrations near the bed, significant amounts of fine-grained sediments have to be resuspended and/or eroded. The disposal grounds of dredged material, navigation channels and adjacent areas with freshly deposited mud have been found to be the major source of the fine-grained sediments during storms. This result is important, as it suggests that dredging and the associated disposal of sediments have made available fine-grained matter that contributes significantly to the formation of high SPM concentrations and high concentrated mud suspensions.

Lenticular Shale Fabrics Resulting from Intermittent Erosion of Water-Rich Muds--Interpreting the Rock Record in the Light of Recent Flume Experiments

Lenticular lamination is a fabric that is known from shales of all ages, but its origin and paleoenvironmental significance is poorly understood. We have successfully reproduced this fabric in flume experiments. Beds of water-rich mud were eroded in a flume and yielded sub-millimeter to centimeter-size fragments that can be transported in bedload for distances of ten kilometers or more. Upon redeposition and compaction, these deposits have the same textural qualities as lenticular laminated shales from the rock record. Although accumulation of fecal pellets or abundant burrow tubes in a shale may produce comparable fabrics upon compaction, these can be distinguished from erosion-produced lenticular lamination via petrographic criteria. Lenticular lamination in shales that is due to deposition of water-rich mud fragments most likely records intermittent erosion and transport of surficial muds by currents.

Tidal sediment transport versus freshwater flood events in the Konkouré Estuary, Republic of Guinea

In comparison to their temperate counterparts, sediment processes in tropical estuaries are poorly known and especially in African ones. The hydrodynamics of such environments is controlled by a combination of multiple processes including morphology, salinity, mangrove vegetation, tidal processes, river discharge, settling and erosion of mud and by physico-chemical processes as well as sediment dynamics. The aim of this study is to understand the sediment processes in this transitional stage of the estuary when the balance between river discharges and marine processes is reversing. Studying the hydrodynamics and sediment dynamics of the Konkouré Estuary has recently been made possible thanks to new data on bathymetry, sedimentary cover, salinity, water elevations, and current velocities. The Lower Konkouré is a shallow, funnel shaped, mesotidal mangrove-fringed, tide-dominated estuary, well mixed during low river discharge and stratified during high river discharge. The Konkouré Estuary is turbid despite the small amount of terrestrial input and its residual velocity at the mouth during low river discharges, landwards for two of the three branches, suggests a landward migration by tidal pumping of the suspended particulate matter. A Turbidity Maximum Zone (TMZ) is identified for typical states of the estuary with regard to fluvial and tidal components. Suspended sediment transport during a transitional stage between the rainy and dry seasons is known thanks to current velocity and Suspended Sediment Concentration (SSC) measurements taken in November 2003. The Richardson layered number calculation assesses that turbulence is the major mixing process in the water column, at least during the flood and ebb stages, whereas stratification occurs during the slack water periods. Tidal currents generate bottom erosion, and turbulence mixes the suspended sediment throughout the water column. As a result, a net sediment input is calculated from the western Konkouré outlet for two consecutive tidal cycles. Despite the net water export, almost 300 tons per tide reach the estuary through this outlet, for a moderate river flow.

Distribution and origin of hybrid beds in sand-rich submarine fans of the Tanqua depocentre, Karoo Basin, South Africa

This study documents the stratigraphic and palaeogeographic distribution of hybrid event beds that comprise both debris-flow (cohesive) and turbidity current (non-cohesive) deposits. This is the first study of such beds in a submarine fan system to combine outcrop and research borehole control, and uses a dataset from the Skoorsteenberg Formation of the Tanqua depocentre in the Karoo Basin, South Africa. Three types of 0.1–1.0 m thick hybrid beds are observed, which have a basal weakly graded fine-grained sandstone turbidite division overlain by a division of variable composition that can comprise 1) poorly sorted carbonaceous-rich material supported by a mud-rich and micaceous sand-matrix; 2) poorly sorted mudstone clasts in a mud-rich sand-silt matrix; or 3) gravel-grade, rounded mudstone clasts in a well sorted (mud-poor) sandstone matrix. These upper divisions are interpreted respectively as: 1) the deposit of a debris-flow most likely derived from shelf-edge collapse; 2) the deposit of a debris flow, most likely developed through flow transformation from turbidity current that eroded a muddy substrate; and 3) from a turbidity current with mudstone clasts transported towards the rear of the flow. All three hybrid bed types are found concentrated at the fringes of lobes that were deposited during fan initiation and growth. The basinward stepping of successive lobes means that the hybrid beds are concentrated at the base of stratigraphic successions in medial and distal fan settings. Hybrid beds are absent in proximal fan positions, and rare and thin in landward-stepping lobes deposited during fan retreat. This distribution is interpreted to reflect the enhanced amounts of erosion and availability of mud along the transport route during early lowstands of sea level. Therefore, hybrid beds can be used to indicate a fan fringe setting, infer lobe stacking patterns, and have a sequence stratigraphic significance.

Trace fossils from Talchir carbonate concretions, Giridih basin, Jharkhand

The carbonate concretions occurring at the bottom of Talchir fissile shale facies preserved signatures of various trace fossils along with a cast of doubtful organisms and cyanobacterial mat structures. The host shale deposited under glacial melt water fed lacustrine condition. The concretions, formed in poorly oxygenated conditions, are either of syndepositional origin and/or deposited a little below the sediment water interface and were later exhumed to the depositional surface due to erosion of soft mud overlying them.The trace fossils are both megascopic and microscopic in nature. The megascopic trace fossils are identified on the basis of their morphology as Monocraterion and Rhizocorallium. Some of the megascopic structures described remain problematic at present. The microscopic trace fossils are formed due to the activity of marine meiofauna (possibly by nematodes), which, although produced morphologically show similar traces of known larger ichnogeneras but much smaller than them. The discovery of these trace fossils apparently indicate the influx of saline water into a lacustrine domain during the Talchir sedimentation at Giridih basin. Moreover, presence of the above two megascopic trace fossils in the marine lacustrine carbonate concretions may lead researchers to consider their much wider environmental significance than hitherto believed.

Erosion d’une vase de l’estuaire de la Loire sous l’action du courant

Erosion d’une vase de l’estuaire de la Loire sous l’action du courant

Modeling a dynamically varying mixed sediment bed with erosion, deposition, bioturbation, consolidation, and armoring

Erosion and deposition of bottom sediments reflect a continual, dynamic adjustment between the fluid forces applied to a sediment bed and the condition of the bed itself. Erosion of fine and mixed sediment beds depends on their composition, their vertical structure, their disturbance/recovery history, and the biota that inhabit them. This paper presents a new one-dimensional (1D), multi-layer sediment bed model for simulating erosion and deposition of fine and mixed sediments subject to consolidation, armoring, and bioturbation. The distinguishing characteristics of this model are a greatly simplified first-order relaxation treatment for consolidation, a mud erosion formulation that adapts to both Type I and II erosion behavior and is based directly on observations, a continuous deposition formulation for mud that can mimic exclusive erosion and deposition behavior, and straightforward inclusion of bioturbation effects. Very good agreement with two laboratory data sets on consolidation effects is achieved by adjusting only the first-order consolidation rate r c . Full model simulations of three idealized cases based on upper Chesapeake Bay, USA observations are presented. In the mud only case, fluid stresses match mud critical stresses at maximum erosion. A consolidation lag results in higher suspended sediment concentrations after erosional events. Erosion occurs only during accelerating currents and deposition does not occur until just before slack water. In the mixed mud and sand case without bioturbation, distinct layers of high and low sand content form and mud suspension is strongly limited by sand armoring. In the mixed mud and sand case with bioturbation, suspended mud concentrations are greater than or equal to either of the other cases. Low surface critical stresses are mixed down into the bed, constrained by the tendency to return towards equilibrium. Sand layers and the potential for armoring of the bed develop briefly, but mix rapidly. This model offers a relatively simple and robust tool for simulating the complex interactions that can affect muddy and mixed sediment bed erodibility.

Erosion and deposition of mud beneath random waves

A systematic approach for predicting the mean erosion and mean deposition rates of mud beneath random waves is derived. This has been accomplished by applying formulas valid for regular waves and by describing the waves as a stationary Gaussian narrow-band random process. The present approach covers flow in the laminar, smooth turbulent and rough turbulent flow regimes. Examples are given, using data typical for field conditions representing laminar and smooth turbulent flow conditions, which are the most common flow regimes over mud beds.

EXPERIMENTAL STUDY ON THE EFFECT OF BIOFILM FOR BED MUD EROSION IN ANNULAR FLUME

The prediction of fine cohesive sediment erosion is important information for civil engineers. However, to calculate an actual amount of erosion, the necessity for considering the effect by biofilm formed on the bed mud surface has been caused. Biofilm formed on the bed mud has a major influence on the erosion phenomenon in estuaries. Therefore, many researchers are grappling from various angles with this problem.Experiments were carried out in an annular flume at the Water and Environmental System Laboratory of Fukuoka University. Results of an experimental study on the effect of biofilm for fine cohesive sediment erosion in the annular flume are described. The relation between biofilm on the bed mud and the initial rate of erosion were conducted under the condition to adjust the amount of light and water temperature. The thickness of biofilm reached about 2mm in the maximum and 0.5mm in the minimum. In this study, it has aimed to clarify how the amount of bed mud erosion changes by the biofilm experimentally

Reservoir Bank Deformation Modeling: Application to Grangent Reservoir

Within inshore or fluvial environments, submerged fine matter mud banks are characterized by a high water content, a great spatial variability, and a strong deformability. The study of their instabilities induced by the variation of hydraulic stress requires a coupled modeling of sliding, erosion, and deposition mechanisms. In order to predict the impact of dam reservoir emptying on the stability of immersed upstream slopes, the method of approach to the problem proposed here combines theoretical developments, numerical modeling, site observations, and measurements. First, the theoretically achieved sliding criterion is compared with unstable mud height measurements. For more accuracy in the representation of the natural events, the sliding criterion is then integrated within a numerical code which couples the computation of hydrodynamic conditions, the erosion, and deposition of mud and the banks sliding. Finally, the results of the combination of all these mechanisms are compared with the variations in the bathymetric profiles obtained on the experimental site.

The influence of Amazon-derived mud banks on the morphology of sandy headland-bound beaches in Cayenne, French Guiana: a short- to long-term perspective

The 1600-km-long South American coast between the Amazon and the Orinoco Deltas has a prograded muddy shoreline comprised of sandy cheniers. The 15-km-long bedrock coast of Cayenne, French Guiana, differs from the rest of this fluctuating alluvial coast in that it comprises several headland-bound fringing sandy barriers that have experienced limited progradation and do not behave as cheniers that migrate landward over a muddy base. The limited progradation of these barriers is probably due to blanketing of fluvial sand by the cover of Amazon mud on the inner shelf, which precludes shoreward reworking to supply the beach and barrier systems. Aerial photographs, satellite images and recent field observations show that alternations between ‘mud bank’ phases, when Amazon-derived mud banks migrating northwestwards attain the Cayenne area, and ‘interbank’ phases, characterised by relative scarcity of mud, result in marked spatial and temporal variations in beach dynamics and morphology. These include periodic alternations in longshore drift that lead to a rare form of total or partial beach rotation that does not result, as is generally reported for nonmud bank-affected beaches, from seasonal variations in deepwater wave approach directions. These beach morphological changes have been defined in terms of a simple, four-stage conceptual model comprising ‘bank’, ‘interbank’ and transitional phases. In extreme ‘bank’ phases, the mud banks responsible for muddy coastal progradation on the Amazon–Orinoco coast may directly weld onto the beaches for periods ranging from months to years, leading to a rare example at the world scale wherein the dynamics of ocean-facing beaches are completely muted under the protective cover of mud that may become rapidly colonised by mangroves. The observations also suggest that these changes do not affect the medium-term (order of tens of years) beach sand budgets. Mud welding onto the beaches may, however, sequester at short time scales (order of months to years) sand eroded from the beaches. Subsequent mud erosion leads to release of this sand and its restitution to the beach sand budget.

Eroded Material Considered a Pollutant

Ross Adams owned 76 acres (31 ha) of land in the North Georgia mountains. David and Barbara Driscoll owned 5 acres (2 ha) adjacent to Adams's property, and Ruel and Patricia Galbreath owned 2 acres (0.8 ha) adjacent to the Driscolls' property. The Spiva Branch stream flows downhill from Adams's property through a pond on the Driscolls' property and then through another pond on the Galbreaths' property before merging with the Nottely River. Adams harvested timber on his property from March to November 1995. During the harvest, he cut and graded roads, installed storm pipes, and cut and removed timber. The harvest and development caused erosion of mud, sand, and other materials on his property. Stormwater washed mud, silt, and other materials from Adams's property into the Spiva Branch stream and then into the Driscolls' and Galbreaths' ponds. Adams failed to seek proper approval from any federal, state, or local government agency before starting to work on his property. The Driscolls and Galbreaths sued Adams for violation of the Clean Water Act (CWA). The trial court ruled against them, concluding that the requirement of a National Pollutant Discharge Elimination System permit was impossible here. On appeal, Adams argued that the discharges did not fall within the scope of prohibited pollutant discharges under the CWA. But, the court said that the definition of “pollutant” in the CWA was broad, including rock, sand, cellar dirt, and industrial, municipal, and agricultural waste. The court also said that when rain water flows from a site on which land‐disturbing activities have been conducted, it falls within the description of a pollutant. The trial court decision was reversed, and the case sent back for further proceedings.

Two-dimensional numerical modeling of storm deposition on the northern California shelf

A two-dimensional, across-shelf sediment-transport model has been developed to simulate storm deposition on continental shelves. This paper describes the structure of the model and related algorithms. In the model, the time-dependent sediment resuspension, transportation and deposition, and across-shelf transport gradients are simulated as responses to storm waves on the Eel shelf, northern California. The simulations show that the pattern of storm deposition on the sandy inner shelf is different from that on the muddy middle and outer shelf. Storm-bed thickness decreases seaward across the inner-shelf sandy zone as far as the 50-m isobath. Farther seaward, storm-bed thickness increases as the mud component in the sea floor becomes abundant, then decreases again, as the weakening of bottom-wave motion with greater depth becomes the dominant control. Storm-bed thickness in the mud zone thus has a nearly symmetrical cross-shelf pattern, with the maximum thickness at a water depth of 70 m. The pattern suggests that while the sequence of new beds found in February 1995 on the Eel shelf may have had a flood source, their geometry reflects storm resuspension and transport. The new muds found in these beds may have been reworked, transported and redeposited several times by storm waves and other oceanic forces. Numerical simulations illustrate that the critical bottom shear stress for mud erosion, τ e, and the coefficient for sand resuspension, γ 0, are more important than other parameters in determining storm depositional patterns on the Eel shelf.

Combating severe fluid erosion and corrosion of drill bits using thermal spray coatings

Thermally sprayed tungsten carbide based coating offers an avenue to minimize severe fluid erosion, wear and corrosion encountered on drill bits and downhole tool assemblies used in mining, oil and gas drilling. The application of high rotary speeds and weights, presence of corrosive elements in the drilling muds, and high velocity mud with entrained cuttings subject these tools and drill bits to severe wear, erosion and corrosion and limit their service life. This paper describes the successful application of thermally sprayed coatings in minimizing the fluid erosion, abrasive wear and corrosion in three cone and polycrystalline diamond bits. The coating processes used are the as-deposited Super D-Gun coatings, spray and fuse with high powered laser and vacuum furnace coatings. The results of mud erosion test and field tests indicate that the detonation gun coated tungsten carbide coatings exhibit superior erosion resistance than all other coatings tested. The hard coatings can be applied on finished drill bit cones and polycrystalline diamond compact bits without damaging the cutters or substrate material.

Biodeposition in a juvenile mussel bed of the east Frisian Wadden Sea (Southern North Sea)

Seasonal trends of biodeposition and animal coverage in backbarrier beds of the musselMytilus edulis in the East Frisian Wadden Sea were documented and quantified to assess the effects of physical disturbance by storms and to develop a conceptual model for the interpretation of biodeposits in the stratigraphic record. Accretion and erosion of biogenic muds (faeces and pseudofaeces) were quantified by repeated precision levelling of a 12 m2 test area. Animal coverage was documented by vertical photography prior to each levelling survey. Levelling and photography were conducted from a bridge and rail system to avoid physical disturbance by trampling. Four surveys, covering a period of different weather conditions, showed variable mean accretion/erosion rates of biodeposits ranging from +0.56 mm d−1 (accretion) to −1.34 mm d−1 (erosion) and areal coverage of mussels ranging from 39.1% to 0%. A pronounced seasonal trend in the biodepositional pattern was observed. Mean accretion rates were high (>0.5 mm d−1) during the summer season, whereas in the winter season net deposition decreased to zero or showed erosional trends. Longterm sediment budgets are thus composed of highly variable mean annual rates which, in turn, comprise highly variable seasonal and even monthly rates, depending on the local weather pattern. This forms the basis for a conceptual model, in terms of which the stratigraphic record of fossil biodeposits can be meaningfully interpreted.

The evolution of the fine sediment regime of the Severn Estuary and Bristol Channel

Fine sediment forms tidal flats along most of the Severn Estuary coastline but subtidal deposits are localized, mainly confined to Newport Deep and Bridgwater Bay. The estuary is experiencing a period of sea level rise which has led to coastal mud erosion in recent centuries. Sediment exchanges with the sea and river inputs are negligible compared to the large-scale exchanges within the estuary itself. Measurements of tidal flat level changes, archaeological finds and anthropogenic chemical concentrations confirm that mud flat erosion is long term and typical of the entire estuary. Sea level rise and coast erosion would normally lead to much of this sediment being redeposited inshore at the landward limit of the alluvium. Other than in narrow salt marsh strips, such deposition is prevented in the Severn by the river walls. In cores of subtidal muddy sediment, the number and thickness of sand layers decreases upwards. The top metre or so is entirely mud and exhibits thousands of layers of varying thickness and origin. Radiochemical analysis confirms that mud is accumulating in the subtidal zone of Newport Deep and the seaward periphery of the Bridgwater Bay mud patch and is increasing in proportion to sand. Circumstantial evidence strongly suggests that for at least the last 600 years the dominant pattern has been one of erosion of fine sediment from the coastal margin and its accumulation in subtidal sinks.