UNCORRECTED PROOF. Bioturbation. Introduction. Particle Reworking. David H Shull, Western Washington University, Bellingham, WA, United States

Transcription

1 David H Shull, Western Washington University, Bellingham, WA, United States address: david. shull@ wwu. edu (D.H. Shull) Introduction 1 Particle Reworking 1 Quantifying Particle Reworking 2 Measuring Particle Reworking 3 Pore Water Bioirrigation 4 and the Ecology and Evolution of Benthic Communities 6 Further Reading 6 Introduction Activities of organisms inhabiting seafloor sediments (termed benthic infauna) are often concealed from view but their effects on sediment chemical and physical properties are nevertheless apparent. Sediment ingestion, the construction of pits, mounds, fecal pellets and burrows, and the ventilation of subsurface burrows with overlying water significantly alter rates of chemical reactions and sedimentwater exchange, destroy signals of stratigraphic tracers, bury reactive organic matter, exhume buried chemical contaminants, and change sediment physical properties such as grain size, porosity, and permeability. Biogenic sediment reworking resulting in a detectable change in sediment physical and chemical properties is termed bioturbation. It is critical to account for bioturbation when calculating chemical fluxes at the sediment water interface and when interpreting sediment stratigraphy and chemical profiles in sediments. In the narrowest sense, bioturbation refers to the biogenic transport of particles that destroys stratigraphic signals. In the broader sense it refers to biogenic transport of pore water and changes in sediment physical properties due to organism activity as well. and its effects on sediment chemistry and stratigraphy is a natural consequence of adaptation by benthic infauna to living and foraging in sediments. Particle Reworking Deposit feeding, the ingestion of particles comprising sedimentary deposits, is the dominant feeding strategy in muddy sediments. In fact, since the vast majority of the ocean is underlain by muddy sediments, deposit feeding is the dominant feeding strategy on the majority of the earth's solid surface. Because digestible organic matter typically comprises less than 1% of sediments by mass, to meet their metabolic needs deposit feeders exhibit rapid sediment ingestion rates, averaging roughly three body weights per day. And, deposit feeders adapted to living in sediments with relatively low organic matter concentrations can exhibit even higher ingestion rates. Rates of deposit feeding of individual organisms increase with increasing body size so that particle reworking rates in some sedimentary deposits may be controlled by a handful of larger species. Deposit feeders employ a wide variety of strategies to collect particles for food; but, reworking modes due to deposit feeding can be broken down into the following categories: conveyor belt feeding where particles are collected at depth and deposited at the sediment surface, subductive feeding, where particles are collected at or near the sediment surface and deposited at depth, and interior feeding where particles are collected and deposited within the sediment column. These feeding modes transport particles the length of the organism's body or the length of its burrow. Some species of deposit feeders also ingest and egest sediment near the sediment surface, resulting in horizontal movement of particles but limited vertical displacement. Due to rapid particle ingestion rates and relatively large horizontal and vertical transport distances, deposit feeding is considered to be the primary agent of particle reworking. But, epibenthic predators including skates, rays, and benthic feeding marine mammals such as gray whales and walrus can also cause extensive reworking in sediment patches where they are feeding. Benthic organisms also rework sediments through burrow formation. Muddy sediments behave more like elastic solids than granular material. A benthic burrower in muddy sediments uses its burrowing apparatus (bivalve foot, polychaete proboscis, amphipod carapace, or Change History: David H Shull updated section and the Ecology and Evolution of Benthic Communities, equations, and modified terminology throughout chapter based on a recent revision by Kristensen et al. (2012) and Modified discussion of assumptions of the biodiffusion coefficient based on the work of Lecroart et al. (2010) and updated Further Reading section. This is an update of D.H. Shull,, Encyclopedia of Ocean Sciences (Second Edition), edited by John H. Steele, Academic Press, 2009, Pages Earth Systems and Environmental Sciences, Volume / B

2 2 other burrowing tool) like a wedge to create and propagate cracks in sediments. After passing through a crack, sediments tend to rebound viscoelastically resulting in relatively little net movement of particles. This generality does not hold in sandy sediments inhabited by taxa such as large irregular urchins that move through sandy sediments like subterranean bulldozers. From a particle's perspective, particle reworking consists of relatively short lived intervals of particle movement due to deposit feeding or burrowing interspersed by relatively long periods during which the particles remain at rest. When particles pass through animal guts, in addition to changing location, the particles may be aggregated into fecal pellets (particles surrounded by or embedded in mucopolymers). When constructing burrows, some infauna produce mucopolymer glue to form sturdy burrow walls, locking particles into place for an extended period of time. Transport of subsurface particles to the sediment surface by conveyor belt feeding results in downward gravitative movement of particles within the sediment column as subsurface feeding voids are filled with sediment from above. Within a particular sedimentary habitat many particle reworking mechanisms occur simultaneously. Quantifying Particle Reworking There are many ways to quantify mathematically the ensemble of particle motions that results in bioturbation. Traditionally, bioturbation has been modeled as though it were analogous to diffusion. This means that the collection of particle motions resembles a large number of small random displacements. Given these assumptions, bioturbation is included in the general diagenetic equation as a biodiffusion coefficient, D B. Whether or not these assumptions are met depends upon the relevant time scale of the tracer of interest and the number of particle transport events that occur over this time scale. Numerous transport events over the course of the relevant time scale generally leads to mixing that resembles diffusion. Ignoring vertical gradients in porosity and sediment compaction, the rate of change of a chemical tracer, C, in the vertical spatial dimension, x, can be represented as follows: where D B is the biodiffusion coefficient (cm year ), u is the sediment accumulation rate (cm year ), and ΣR represents the sum of chemical reactions. In the absence of specific information on particle reworking mechanisms, it is often assumed that D B is constant throughout the reworked layer to the depth L. Below the depth L, D B is zero. The advantage of this formulation is that all the various particle reworking processes are quantified by one parameter, D B. The non dimensional Peclet number, Pe = uld B 1, quantifies the relative importance of bioturbation versus sediment accumulation in determining the profile of C within the reworked layer. Values of Pe less than one indicate stronger influence of bioturbation. Table 1 summarizes the general pattern of variation in D B, u, L, and Pe among benthic provinces at different water depths. The depth of the reworked layer, L, shows little systematic variation among habitats, averaging 10 cm. Although there is considerable variation in Pe, the low values in each province indicate that bioturbation is generally important throughout the ocean. An easy to remember rule of thumb regarding bioturbation rates is that D B varies from approximately to 100 cm year from deep sea to shallow water sediments. This variation is correlated with increased abundance of infauna, greater rates of food supply, and (with the exception of polar regions) elevated average bottom water temperatures with decreasing water depth. Because particle reworking mechanisms can transport particles relatively large distances, roughly the length of the reworked zone, L, and because particle trajectories are often non random, the biodiffusion coefficient might not be appropriate for modeling the effects of bioturbation on transport of some tracers. A more general model of particle mixing that includes longer distance, non random particle trajectories is the non local bioturbation model. Again neglecting variation in porosity: Table 1 Variation in the biodiffusion coefficient, D B, sedimentation rate, u, and the Peclet number, Pe, characteristic of different benthic environments Region D (cm year ) u (cm y ) L (cm) Pe B Shallow water Continental shelf Continental slope Deep sea A Peclet number less than one indicates bioturbation is more important for transport relative to sedimentation. A Peclet number greater than one indicates sediment accumulation is more important than bioturbation. (1)

3 3 (2) 1 where K is the exchange function (dimensions: time ) that quantifies the rate of particle movement from one depth, x, to any other depth, x. The first term on the right hand side gives the concentration change at depth x due to the delivery of a particle tracer from other depths, x. The second term gives the concentration change at depth x due to transport of a tracer from depth x to other depths x. The other terms are defined as in Eq. (1). The exchange function, K, can potentially quantify a complex ensemble of particle reworking mechanisms. Analogues of Eq. (2) that rely on discrete mathematics exist. In one dimension, non local transport can be modeled as a transition matrix in which the rows of the transition matrix correspond to depths in the sediment where particles are collected (x), the columns correspond to the depths where particles are deposited (x ) and the matrix elements quantify the probability of transport of a tracer among depths. Multiple dimensional automaton models can simulate complex modes of particle transport in both vertical and horizontal dimensions. These more complex models can better capture the complexities of particle reworking but sacrifice the one parameter simplicity of Eq. (1). Measuring Particle Reworking There are two common approaches for determining the values of the bioturbation parameters in these models. Mixing parameters can be estimated from direct measurements of deposit feeder ingestion rates and organism burrowing rates. But more often these parameters are estimated by measuring sediment bound tracers with known inputs to the sediment and known reaction rates (ΣR). Mixing parameters are then calculated by fitting measured tracer profiles to profiles calculated by use of the appropriate mathematical model. Useful particle reworking tracers include excess activities of naturally occurring particle reactive radionuclides such as Th, Pb, or Be. These radionuclides have a relatively continuous source either from the atmosphere or from the overlying water column, are rapidly scavenged onto particles and sink to the sea floor where they are mixed into the sediment bed. In addition, chlorophyll a, artificial tracers added to the sediment surface as an impulse such as glass beads or fluorescent luminophores, or other exotic identifiable material with a known rate of input are used as tracers of particle reworking. The profile of excess Pb in Fig. 1 illustrates several effects of bioturbation on a tracer profile. Rates of bioturbation in the top 6 cm are rapid enough to completely mix excess Pb within this layer. The subsurface peak at cm indicates subsurface deposition of surficial material. Below 16 cm, the slope of the profile is set by the rate of sediment accumulation and radioactive decay of Pb (half life = 22.3 year). Particle reworking has important consequences for sediment stratigraphy, chemistry and biology. Particle reworking can homogenize tracers within the reworked layer (Fig. 1). It acts as a low pass filter, destroying information deposited on short time scales, but preserving 3 longer term trends. makes it generally difficult or impossible to resolve time scales of less than 10 year stratigraphically in deep sea sediment cores. If the particle reworking mechanism is not known, it is difficult to separate changes in the input signal from changes due to mixing (Fig. 2). If mixing is not complete, and the particle reworking mechanism is known, it may be possible to Fig. 1 Excess Pb activity versus depth in a sediment core from Narragansett Bay, Rhode Island. Data from Shull, D. H. Transition matrix model of bioturbation and radionuclide diagenesis. Limnology and Oceanography, vol. 46, pp , copyright 2001, with permission from the American Society of Limnology and Oceanography.

4 4 Fig. 2 Changes in the profile of a hypothetical conservative tracer present initially as two narrow subsurface peaks, as predicted from Eq. (1). D = 0.1 cm 2 year 1, u = 0.1 cm year 1, L = 10 cm. Solid line: Tracer profile at t = 0 year. Dotted line: t = 25 year. Dashed line: t = 150 year. B deconvolve the input signal to the stratigraphic record, although detailed information will be lost. If bioturbation in the surface reworked zone completely homogenizes a tracer, then knowing the mixing mechanism is unimportant. Once pancake batter is thoroughly mixed, for example, it no longer contains information on how the mixing was performed. Particle reworking has important consequences for sediment geochemistry. It buries reactive organic matter. Subductive deposit feeders selectively feed on organic rich particles near the sediment surface and deposit them at depth, perhaps as food caches. In the presence of horizontal transport of organic matter, or suspension feeding benthos that locally enhance organic matter deposition through biodeposit formation, particle reworking can greatly enhance the organic matter inventory in sediments. Burial of organic matter exposes it to different oxidants, changing the degradation pathway. In particular, reworking of Mn and Fe oxides cycles them between oxidative and reducing environments, resulting in enhanced anaerobic degradation of organic matter. Particle reworking changes sediment properties as well. Pelletization changes the sediment grain size distribution. Furthermore, particle reworking rates are particle size dependent. Size selective feeding by deposit feeders results in biogenic graded bedding with lag layers of large sediment particles either at the sediment surface or at depth, depending upon the particle reworking mechanism. Formation of pellets and burrows increases sediment porosity, counteracting the effects of sediment compaction. Sediment surface manifestations of particle reworking such as pits, mounds, and tubes alter seafloor roughness and flow characteristics of the benthic boundary layer roughly doubling the drag compared to a hydrodynamically smooth bed. Pore Water Bioirrigation Most benthic infauna maintain a burrow that connects to the sediment water interface to facilitate respiration, feeding, defecation and other metabolic processes. These burrows exist in a range of geometries including vertical cylinders, U or J shaped tubes, or branching networks. Because dissolved oxygen in nearshore sediments often penetrates no farther than a few millimeters, benthic infauna ventilate their burrows to meet their metabolic requirements for oxygen. They accomplish this by thrashing their bodies, flapping appendages, by peristalsis, by beating cilia, or by oscillating like pistons in their tubes. These ventilation mechanisms result in intermittent burrow flushing, which exchanges a portion of the fluid inside the burrow with overlying water. In this way, organisms in the sediment can flush out metabolic wastes and toxins such as hydrogen sulfide that have accumulated in their burrows and they can restock the burrow water with dissolved oxygen. Observations of organisms in artificial tubes maintained in the laboratory indicate that burrow ventilation is episodic, with ventilation frequencies ranging from once per hour to ten or more ventilation events per hour. Deposit feeding infauna generally ventilate less frequently than suspension feeding infauna, which pump water through their burrows for both respiration and food capture. The sediments into which these burrows are built are mixtures of particles and interconnected pore water. Surficial sediments may possess porosities (defined as the volume of interconnected pore water per unit volume of sediment) in excess of 90%. Thus, surface sediments generally contain more pore water than particles. The rate of molecular diffusion of solutes through pore water is reduced relative to diffusion in a free solution because the solutes must follow a winding path through the particles, a property called sediment tortuosity. Particle reworking redistributes this pore water along with the particles, but since rates of pore water transport, inferred from dissolved tracer distributions, are typically an order of magnitude higher than rates of particle reworking, particle reworking is a relatively unimportant mechanism for transporting pore water. Rather, burrow ventilation seems the most important biogenic mechanism of porewater transport. The consequences of burrow ventilation on pore water transport in the surrounding sediments (termed bioirrigation) depend upon sediment permeability.

5 5 Sandy sediment typically possesses high enough permeability to allow advective flow of pore water through the interconnected pore space surrounding sediment particles. Under these conditions, the pressure head within a burrow created by burrow ventilation activities can drive pore water flow from the burrow into surrounding sediments. The velocity of this flow can be calculated using Darcy's Law: where u d is the Darcy velocity, k is sediment permeability, μ is the dynamic viscosity of pore water, ρ is the pore water density, g is gravity, and Δ is a gradient operator (e.g., / x, / y). The velocity of pore water, u is related to the Darcy velocity, u d = uφ, where φ = porosity. Substituting Eq. (3) into the general diagenetic equation gives the expected change in concentration of a pore water tracer subjected to an advection velocity driven by burrow ventilation, molecular diffusion and chemical reactions. where D M is the molecular diffusion coefficient, corrected for tortuosity. In contrast to sandy sediments, permeability of mud is generally too low to permit significant pore water advection so that u = 0. Thus, pore water transport in muddy sediments is dominated by molecular diffusion. Burrow ventilation in muddy sediments enhances porewater transport by changing the diffusive geometry. Fig. 3 shows the geometry of idealized equally spaced vertical burrows embedded into sediment. If these burrows were rapidly flushed so that the solute concentration within the burrows were equal to the solute concentration in the overlying water, then, the corresponding diagenetic equation governing pore water transport in the vertical dimension, x, and in the radial dimension, r, would be given by The diffusion operator within the parentheses is similar to the diffusion operator in Eq. (4), but quantifies molecular diffusion in both the vertical (x) and radial (r) dimensions. A one dimensional diagenetic model that incorporates the effects of bioirrigation on pore water transport can be derived from Eq. (5): 1 where α(d ) is the coefficient of non local bioirrigation, and C 0 is the concentration of the solute tracer in overlying water. The nonlocal bioirrigation coefficient, α, in Eq. (6) treats bioirrigation as both a source and a sink for solutes at depth. The value of the bioirrigation exchange rate, α, is usually determined by measuring dissolved pore water tracers with known inputs 222 and reaction kinetics. The most commonly used radionuclide tracer of bioirrigation is Rn. Produced within sediments from the decay of its parent Ra, Rn is a soluble noble gas. Pore water exchange with overlying water results in lower Rn activity in sediment pore waters than would be expected, compared to the activity of its parent Ra. The shape of the Rn profile and the magnitude of the Rn 226 deficit relative to Ra are used to determine rates of bioirrigation (Fig. 4). Other tracers of pore water exchange include inert solutes (3) (4) (5) (6) Fig. 3 Idealized burrow geometry underlying Eq. (5). (A) Burrows as close packed cylinders. (B) Rectangular plane intersecting an average burrow microenvironment. (C) The x r domain of Eq. (5). The shaded rectangle represents the burrow, the unshaded rectangle represents surrounding sediment.

6 6 Fig. 4 Measured 222Rn activity and supported 222Rn activity (produced from the decay of 226Ra within sediment particles) versus depth in a sediment core from Boston Harbor, Massachusetts. Horizontal error bars represent standard error from three replicate cores. The bioirrigation rate, α (d 1), was modeled as the exponential function. α = 3.8 e x, and the modeled profile was calculated from Eq. (6). Data from Shull, D. H., previously unpublished. such as bromide introduced to the sediment in incubation experiments or dissolved nutrients such as ammonium, nitrate or silicate, if reaction kinetics can be estimated. Bioirrigation has important implications for sediment geochemistry. Bioirrigation accelerates sediment water fluxes, changes rates of elemental cycling, catalyzes oxidation reactions in the sediment, changes vertical and horizontal gradients of pore water solutes, elevates levels of dissolved oxygen and reduces exposure of organisms to metabolic wastes. The increased transport of dissolved oxygen within the sediment, enlarged surface area for pore water exchange within the sediment (e.g., across burrow walls) and stimulation of microbial metabolism lead to enhanced sedimentary oxygen consumption. This change in the redox geometry of sediments can significantly alter rates of redox sensitive reactions that occur in sediments such as nitrification, denitrification, sulfate reduction, and mercury methylation. and the Ecology and Evolution of Benthic Communities has numerous effects on benthic community structure. In muddy sediments, particle reworking by deposit feeders appears to reduce densities of suspension feeders. These effects can be species specific. For example, thalassinidean shrimp have particularly strong negative effects on bivalves, which they can smother with resuspended sediment. Conveyor belt feeding can displace surface dwelling benthos. changes the depth distribution of organic matter and can increase the inventory and quality of food for deposit feeders in sediments. The provision of both food and energetically favorable electron acceptors such as oxygen, nitrate, and iron oxide to depth in the sediment leads to higher abundance of bacteria relative to archaea in surficial sediments. Bioirrigation increases microbial activity at depth and in the vicinity of burrows. can also increase nutrient fluxes leading to elevated rates of benthic primary production and increased microbial productivity as well. On the other hand, bioirrigation also enhances rates of denitrification, reducing the efficiency of sedimentary nitrogen cycling. Marine benthic habitats of the late Neoproterozoic ( million years ago) were very different from benthic habitats that existed later in the Phanerozoic. Seafloors at this time were characterized by well developed microbial mats, as suggested by studies of sedimentary fabric preserved in the geologic record. These extensive microbial mats and associated seafloor fauna, such as immobile suspension feeding helicopacoid echinoderms, became scarce or extinct in the Cambrian. The substantial change that occurred in seafloor communities around this time, termed the Cambrian Substrate Revolution is thought to be caused by the development of bioturbation. It is hypothesized that the emergence of both bioturbation and predation around this time led to the extinction of non burrowing taxa and influenced subsequent development of animal body plans during the Cambrian. also made a new food resource, buried organic matter, accessible to deposit feeders while radically changing the geochemistry of the seafloor. Further Reading Aller, R.C., The effects of macrobenthos on chemical properties of marine sediments and overlying waters. In: McCall, P.L., Tevesz, M.J.S. (Eds.), Animal sediment relations. Plenum Publishing, pp Boudreau, B.P., Diagenetic models and their implementation. Springer Verlag. Boudreau, B.P., Jorgensen, B.B., The benthic boundary layer: Transport processes and biogeochemistry. Oxford University Press. Burdige, D.J., Geochemistry of marine sediments. Princeton University Press.

7 7 Dorgan, K.M., Jumars, P.A., Boudreau, B.P., Johnson, B.D., Macrofaunal burrowing: The medium is the message. Oceanography and Marine Biology: An Annual Review 44, Kristensen, E., Penha Lopes, G., Delefosse, M., Valdemarsen, T., Quintana, C.O., Banta, G.T., What is bioturbation? The need for a precise definition for fauna in aquatic sciences. Marine Ecology Progress Series 446, Lecroart, P., Maire, O., Schmidt, S., Grémare, A., Anschutz, P., Meysman, F.J., 2010., short lived radioisotopes, and the tracer dependence of biodiffusion coefficients. Geochimica et Cosmochimica Acta 74, Lohrer, A.M., Thrush, S.F., Gibbs, M.M., Bioturbators enhance ecosystem function through complex biogeochemical interactions. Nature Meysman, F., Boudreau, B.P., Middelburg, J.J., Relations between local, non local, discrete and continuous models of bioturbation. Journal of Marine Research 61,