Biogeochemical consequences of macrofauna burrow ventilation†
© The Royal Society of Chemistry and the Division of Geochemistry of the American Chemical Society 2001
Received: 17 September 2001
Accepted: 08 October 2001
Published: 19 October 2001
The burrow walls created by macrofauna in aquatic sediments are sites of intense chemical mass transfer. Quantitative measurement of their significance is, however, difficult because chemistry in the immediate vicinity of burrow walls is temporally dynamic due to periodic ventilation of burrows by macrofauna. A temporally dynamic, 2D multicomponent diffusion-reaction model was utilized to depict the magnitude and time dependency of chemical mass transfer in the immediate vicinity of burrow walls as well as at the water/sediment interface. The simulation results illustrate that sediment particles, pore water, and microorganisms within a few millimeters of burrow walls experience significant oscillation in pH (as much as two pH units) and dissolved oxygen concentration (between saturation and near anoxia) whereas such oscillation is absent at the water/ sediment interface. The geochemical oscillation is expected to affect the net stability of mineral phases, activities and community structures of microorganisms, and rates and magnitudes of microbial diagenetic reactions.
Burrowing infauna in aquatic sediments induce temporal fluctuation in concentrations of dissolved species (i.e., geochemical oscillation) through their metabolism and burrow ventilation activities. They periodically irrigate their burrows to replace metabolite-rich burrow water with fresh overlying water. The immediate vicinity of burrow walls is subject to periodic changes in the concentrations of oxygen, nutrients, and other pore water species.[1, 2]
Geochemical oscillation affects the courses of organic carbon (OC) diagenesis. Field and laboratory studies suggest that OC remineralization rates are enhanced by the redox oscillation, even though the anoxic period in each oscillation cycle is typically much longer (~10–100X) than the oxic period for a given sediment microenvironment.[4, 5] Temporally averaged redox conditions do not directly correlate to the rates and magnitudes of OC reactions.[3, 6] A comprehensive understanding of OC diagenesis in aquatic sediments thus requires an adequate characterization of the temporal dynamics of redox and other geochemical parameters.
Directly observed data for geochemical oscillation induced by infauna are not abundant. A few studies have measured temporal dynamics of geochemical variables in biologically reworked sediments using redox potential, Eh, and oxygen microprobes.[1, 7, 8] The number of such studies is small despite the importance of geochemical oscillation: this may be largely due to the technical difficulties in measuring the geochemical phenomena that are not only temporally variable but also are confined to small regions. For example, in shallow marine and estuarine sediments, O2 penetration in the vicinity of burrow walls is limited to a few millimeters, and thus oscillation of redox and other related parameters may be restricted to the regions within a few millimeters of burrow walls. The temporal scale of oscillation at burrow walls depends on the burrow ventilation habits of infauna, and each oscillation cycle may be as short as several minutes. Sampling devices typically used by geochemists (e.g., cores and pore water peepers) are designed to collect data that are averaged over the spatial and temporal extents too large to capture such variability. Microelectrode arrays and gel probes have elucidated sub-millimeter spatial variability of geochemical parameters, [11–13] but they have not been extensively used to capture the temporal variability.
Although the direct observations of geochemical oscillation in burrowed sediments are still few, such data can be put into the context of overall diagenesis when assisted with modeling. Recent studies have shown that the computational simulation of reaction couplings and kinetics associated with OC remineralization is a viable technique for the study of OC diagenesis.[14–16] Marinelli and Boudreau applied this type of model to calculate the temporal fluctuation of O2 and pH in the vicinity of model burrows artificially flushed with overlying water, and to estimate the effect of such oscillation on net chemical mass transfer.
In the present study, a numerical model of OC diagenesis and solute transport is used to investigate the magnitude of geochemical oscillation in the vicinity of burrows resulting from macrofauna ventilation and metabolite excretion. The model provides a tool for quantifying the variability of geochemical parameters that change rapidly within a small spatial extent.
The 2D cylinder geometry used in this study is taken from the single-component reaction, diffusion and burrow irrigation model first introduced by Aller. The original model has since been expanded to a multi-solute numerical model for reaction, diffusion and discontinuous burrow irrigation.[11, 18] The model geometry has also been adapted to allow burrows with variable depths. The 2D cylinder geometry is advantageous to 1D vertical expression in describing temporal and spatial variability of geochemical parameters because it explicitly considers the radial chemical mass transfer in the vicinity of burrow walls. Whereas well-constrained 1D steady-state models are able to accurately hindcast net chemical mass transfer with the use of adjustable parameters, they cannot represent radial variability such as the spatial and temporal variability seen in the immediate vicinity of burrow walls.
where ϕ ≡ porosity, D ≡ diffusion coefficient of the solute, θ ≡ diffusive tortuosity, and R ≡ net reaction rate. The diffusive tortuosity term in the equation can be replaced by a porosity expression using the following empirical correlation.
θ2 = 1 - ln (ϕ2) (2)
Reactions and rate expressions considered in the model simulations. Rate symbols are defined in Table 2
Primary redox reactions-
(CH2O) x (NH3) y (H3PO4) z + (x + 2y)O2 + (y + 2z)HCO3- → (x + y + 2z)CO2 + yNO3- + zHPO42- + (x + 2y + 2z)H2O
(CH2O) x (NH3) y (H3PO4) z + ((4x + 3y)/5)NO3- → ((2x + 4y)/5)N2 + ((x - 3y + 10x)/5)CO2 + ((4x + 3y - 10z)/5)HCO3-+ zHPO42- + ((3x + 6y + 10z)/5)H2O
(CH2O) x (NH3) y (H3PO4) z +(x/2)SO42- + (y - 2z)CO2 + (y - 2z)H2O → (x/2)H2S + (x + y - 2z)HCO3- + yNH4+ + zHPO42-
Secondary redox reactions-
NH4+ + 2O2 + 2HCO3- → NO3- + 2CO2 + 3H2O
H2S + 2O2 + 2HCO3- → SO42- + 2CO2 + 2H2O
Acid-base reactions (equilibrium)-
CO2 + H2O → HCO3- + H+
HCO3 ↔ CO32- + H+
H2S ↔ HS- + H+
Adsorption reations (equilibrium)-
NH4+(aq) ↔ NH4+(ads)
Boundary conditions needed for the numerical solution of eqn. (1) include the following:
Cx = 0,r,t= C0 (3)
Solute concentrations at WSI (i.e., x = 0) are assumed to be constant (C0) (eqn. (3)). Solute concentrations of the burrow water during the ventilation period are also assumed equal to the overlying water (eqn. (4)), whereas, during the rest period, the burrow water composition is determined according to eqn. (1) by diffusive transport across the burrow wall, macrofaunal metabolite production, and aerobic reoxidation of reduced species. In reality, burrow water composition does not remain identical to that of the overlying water during ventilation because the volume of water replaced by each pumping of macrofauna is smaller than the volume of burrow cavity and the velocity of water due to macrofauna flushing has a finite value [22, 23]. Radial solute flux at the outer surface of each cylinder (i.e., r = r2) and at burrow axis (i.e., r = 0) are set to zero due to symmetry (eqn. (5) and (6)). Vertical solute flux at the bottom boundary (at x = L) is set to zero (eqn. (7)). The model geometry is based on the assumption that all burrows have the depth extent of L. In actual simulations, L has to be set greater than the depth extent of one's interest in order to avoid numerical artifacts. This is obviously unrealistic because actual burrows have variable vertical extents and some of them may be much shorter than L. Consequently, the model is expected to be the most accurate near WSI: its accuracy decreases with depth. At the burrow wall (i.e., r = r1), solute concentrations and radial flux are both continuous (eqn. (8) and (9)).
A FORTRAN code is written to numerically solve the conservation equation (1) with both left and right hand sides of the equation being fully discretized. A 2D uneven grid is utilized in order to have small grid spacings near WSI and burrow wall (Δx = Δr = 0.25 × 10-3 m) where rapid aerobic reactions are expected to yield steep spatial gradients in concentrations. The Cx,rfor each grid point is solved as a time-evolution problem until the evolution of Cx,rdistribution during a given ventilation-rest cycle becomes identical to the evolution during the previous ventilation-rest cycle.
The conservation equation is written for each of the following species: O2, NO3-, SO42-, NH4+, ΣS (≡ H2S + HS-), ΣCO2 (≡ CO2 + H2CO3 + HCO3- + CO32-), and titration alkalinity (Alkt = HCO3- + 2CO32- + HS-). All conservation equations are coupled and solved simultaneously at each time step through the reaction terms, R, as shown in Table 1. For example, the reaction term for the conservation equation of O2 at time step T is determined by the concentration values of O2, NH4+, and ΣS at time step T – 1 (see eqn. (I-1)).
The diffusion coefficients of solute species, Monod kinetics constants, and thermodynamic constants for sulfide and carbonate systems can be taken from previously established equations found in literature that express these parameters as functions of temperature and salinity. The other site-and species-specific parameters, including the parameters for reaction kinetics and burrow geometry, need to be taken from each environment being studied. The durations of ventilation and rest periods are also required to implement the periodic, discontinuous irrigation scenario.
Directly determined data on macrofauna burrow geometry, ventilation habits, OC degradation rates, porosity, and pore water chemistry are necessary in order to properly constrain the model and evaluate the simulation results. Collective studies of sediments populated with Nereis diversicolor [25–28] provide many of the necessary data.
Model system description
Model simulations are carried out for the mesocosm systems of Kristensen and Hansen. All necessary parameters were given in Kristensen and Hansen or the references therein.[26–28] The mesocosms were loaded with homogenized fine-grained mud with 75% porosity to the depth of 15 cm, and populated with 1200 m-2 Nereis diversicolor of 200–400 mg wet weight. Each N. diversicolor constructed a U-shape burrow with two openings at WSI. The upper 5–6 cm of the sediment columns were burrowed. The ratio for the production of ΣCO2 and NH4+ due to microbial remineralization was given by Kristensen and Hansen to be 4.8, which they determined by correlating the pore water ΣCO2 and NH4+ profiles using the method previously described. These values are much lower than the C: N ratio of bulk organic matter in the source mud from Kertinge Nor, Denmark (≈ 12) given by Hansen and Kristensen because organic matter that is readily decomposed by the microbial remineralization has a lower C: N ratio than the bulk organic matter pool. The bottom water ΣCO2 and NH4+ concentrations were measured to be 1.82 × 10-3 M and 0.014 × 10-3 M, respectively. Kristensen and Hansen also described control mesocosms that were loaded with the same mud but with no macrofauna.
The anaerobic OC degradation rate, ROCan, was determined by fitting the 1D steady-state model to measured ΣCO2 and NH4+ depth profiles of control mesocosms. The control mesocosms were not burrowed, thus the model used for this simulation was 1D, rather than 2D, and assumed steady state:
Parameter values for 1D control mesocosms
Simulation depth a
Bottom water O2 b
0.228 × 10-3
Bottom water NO3-b
0.015 × 10-3
Bottom water SO42-c
Bottom water NH4+a
0.014 × 10-3
Bottom water ΣS b
Bottom water ΣCO2 a
1.820 × 10-3
Bottom water H+b
8.97 × 10-9
O2 diffusion coefficient d
1.82 × 10-9
NO3-diffusion coefficient d
1.53 × 10-9
SO42- diffusion coefficient d
0.84 × 10-9
NH4+ diffusion coefficient d
1.57 × 10-9
H2S diffusion coefficient d
1.36 × 10-9
HS- diffusion coefficient d
1.45 × 10-9
CO2 diffusion coefficient d
1.44 × 10-9
HCO3- diffusion coefficient d
0.92 × 10-9
CO32- diffusion coefficient d
0.73 × 10-9
Aerobic OC degradation rate e
6.9 × 10-8
Anaerobic OC degradation rate f
2.5 × 10-9
C: N ratio a
Aerobic NH4+ reoxidation rate g
O2 Monod sat./inhib. Const. h
0.02 × 10-3
Aerobic ΣS reoxidation rate g
5.07 × 10-3
NO3 Monad sat sat./inhib. const. h
0.005 × 10-3
SO42- Monad sat. const. h
1.6 × 10-3
9.00 × 10-7
4.62 × 10-10
ΣS acidity constant i
1.82 × 10-7
2D simulation results
Effect of ventilation patterns
During the rest period, the burrow water becomes less oxygenated, and more enriched in metabolites such as NH4+ and ΣCO2. The build up of ΣCO2 also induces decrease in pH values. The above calculations imply that the duration of the rest period is an important variable in determining the magnitudes of geochemical oscillations. The longer the rest period, the lower the pH values and O2 concentrations would become, and thus the ranges of oscillation during each rest–ventilation cycle would become greater. In order to illustrate the significance of these variables, the 2D model was applied to a model system that consists of the same mud as the original mesocosms of Kristensen and Hansen but is populated with 1200 m-2 N. virens instead of N. diversicolor which has the rest period of trest = 2391 s and ventilation period of tvent = 840 s, according to a mesocosm study by Kristensen and others. This rest period is nine times greater than the rest period of N. diversicolor.
N. virens construct U-shaped burrows. The model burrow diameter (r1) was determined from the mesocosm study to be r1 = 3.0 × 10-3 m. The same study also measured the NH4+ excretion rate of N. virens to be 2.5 nmol [cm3 (burrow water)-1 min-1 or 4.2 × 10-8 M s-1]. The C:N ratio of 4.8 yields the ΣCO2 excretion rate of 2.0 × 10-7 M s-1.
Difference between two types of interface
The simulations indicate that both WSI and burrow walls are the sites of intense chemical mass transfer as evidenced by the steep gradients in geochemical parameters such as pH and O2 concentrations. Burrowed sediments accommodate greater extent of aerobic OC remineralization than non-bioturbated sediments due to the increased interfacial area between anoxic sediments and oxygenated water. There is, however, one important difference between the redox interface at WSI and the redox interface at burrow walls. The burrow walls are temporally dynamic because they border burrow water whose geochemical properties oscillate according to the metabolism and ventilation of burrowing macrofauna. On the other hand, WSI is adjacent to the overlying water whose geochemical properties are temporally more stable. Consequently, sediment particles and microorganisms in the close vicinity of burrow walls experience temporal oscillation in redox chemistry and pH whereas sediment particles and microorganisms in the immediate vicinity of WSI do not undergo such oscillation. For example, microorganisms in the close vicinity of N. virens burrows have to adapt to the oscillating redox environment in which they may need to switch between different terminal electron acceptors for respiration during a short period of time. In addition, the same microorganisms are required to tolerate the oscillation in pH, which has been shown to influence the growth rates and activities of bacteria. The mineral particles near the burrow walls also experience the same pH oscillation. If mineral particles such as calcareous tests of microorganisms or iron oxyhydroxides are located in the immediate vicinity of burrow walls, they may experience oscillation between supersaturation and undersaturation due to the fluctuation in pH.
These two different types of interfaces (WSI and burrow walls) are expected to accommodate different microbial ecology and distribution of functional groups because their redox, nutrients, and pH environments are distinctively different (i.e., steady state vs. periodic oscillation). While the enhancement of net microbial activities due to the presence of burrows is well documented,[34, 35] the mechanisms for enhancement have not been fully evaluated. Potential enhancement mechanisms include the faster removal of toxic metabolites, increased diffusive transport, increased introduction of terminal electron acceptors, and resulting stimulation of production. Further investigations of the spatial and temporal variability of microbial activities, community structures, and geochemical environments in the immediate vicinity of burrow walls are warranted.
Numerical simulations using a 2D diffusion–reaction model were conducted to examine the geochemical effect of macrofaunal burrow ventilation in the immediate vicinity of burrow walls and at the water/sediment interface. During the rest period, macrofaunal metabolites such as NH4+ and ΣCO2 build up while loss of O2 occurs due to molecular diffusion and aerobic reoxidation of reduced species within the burrow cavity water. On the other hand, the geochemical composition of burrow cavity water becomes rich in O2 and depleted in metabolites during the ventilation periods. Consequently, the redox and other geochemical parameters, including pH, oscillate in the immediate vicinity of burrow walls during each rest–ventilation cycle. The range of oscillation is greater when the burrow occupant maintains longer rest periods. Such oscillation is absent in the vicinity of the water/sediment interface, which borders the overlying water whose composition remains constant.
The redox and pH oscillations depicted through the model simulations are likely to influence the activity and community structure of sedimentary microorganisms and thermodynamic stability of mineral particles. Further studies are necessary in order to quantify the couplings between geochemical variables and microbial properties in temporally dynamic environments such as the vicinity of burrow walls.
† Presented during the ACS Division of Geochemistry symposium 'Biogeochemical Consequences of Dynamic Interactions Between Benthic Fauna, Microbes and Aquatic Sediments', San Diego, April 2001.
This study was originally presented in April, 2001 at ACS National Meeting in San Diego, CA. The ACS symposium titled "Biogeochemical Consequences of Dynamic Interactions Between Benthic Fauna, Microbes, and Aquatic Sediments" was funded in part by ONR342, ACS-PRF, and ACS Geochemistry Division. The study was supported by ONR 322GG (PE No. 0601153N) and NRL Core Funding. NRL contribution number JA/7430-01-0010.
- Forster S: PSZN I: Mar Ecol. 1996, 17: 309-Google Scholar
- Kristensen E: Mar Biol. 1989, 101: 381-10.1007/BF00428134.View ArticleGoogle Scholar
- Aller RC: Chem Geol. 1994, 114: 331-10.1016/0009-2541(94)90062-0.View ArticleGoogle Scholar
- Burdige DJ: Org Geochem. 2001, 32: 487-10.1016/S0146-6380(00)00191-1.View ArticleGoogle Scholar
- Kristensen E, Holmer M: Geochim Cosmochim Acta. 2001, 65: 419-10.1016/S0016-7037(00)00532-9.View ArticleGoogle Scholar
- Hulthe G, Hulth S, Hall POJ: Geochim Cosmochim Acta. 1998, 62: 1319-10.1016/S0016-7037(98)00044-1.View ArticleGoogle Scholar
- Ziebis W, Forster S, Huettel M, Jørgensen BB: Nature. 1996, 382: 619-10.1038/382619a0.View ArticleGoogle Scholar
- Forster S, Graf G: Hydrobiologia. 1992, 235/236: 527-10.1007/BF00026240.View ArticleGoogle Scholar
- Jørgensen BB, Revsbech NP: Limnol Oceanogr. 1985, 30: 111-View ArticleGoogle Scholar
- Kristensen E: Hydrobiologia. 2000, 426: 1-10.1023/A:1003980226194.View ArticleGoogle Scholar
- Marinelli RL, Boudreau BP: J Mar Res. 1996, 54: 939-10.1357/0022240963213646.View ArticleGoogle Scholar
- Luther GW, Brendel PJ, Lewis BL, Sundby B, Lefrancois L, Silverberg N, Nuzzio DB: Limn Oceanogr. 1998, 43: 325-View ArticleGoogle Scholar
- Mortimer RJG, Krom MD, Hall POJ, Hulth S, Ståhl H: Mar Chem. 1998, 63: 119-10.1016/S0304-4203(98)00055-3.View ArticleGoogle Scholar
- Boudreau BP: Comput Geosci. 1996, 22: 479-10.1016/0098-3004(95)00115-8.View ArticleGoogle Scholar
- Soetaert K, Herman PMJ, Middelburg JJ: Geochim Cosmochim Acta. 1996, 60: 1019-10.1016/0016-7037(96)00013-0.View ArticleGoogle Scholar
- Van Cappellen P, Wang Y: Am J Sci. 1996, 296: 197-View ArticleGoogle Scholar
- Aller RC: Geochim Cosmochim Acta. 1980, 44: 1955-10.1016/0016-7037(80)90195-7.View ArticleGoogle Scholar
- Boudreau BP, Marinelli RL: J Mar Res. 1994, 52: 947-10.1357/0022240943076902.View ArticleGoogle Scholar
- Furukawa Y, Bentley SJ, Lavoie DL: J Mar Res. 59: 417-10.1357/002224001762842262.Google Scholar
- Berner RA: Early Diagenesis: A Mathematical Approach. 1980, Princeton University Press, Princeton, NJGoogle Scholar
- Boudreau BP: Diagenetic Models and Their Implementations. 1997, Springer-Verlag, HeidelbergView ArticleGoogle Scholar
- Kristensen E, Jensen MH, Aller RC: J Mar Res. 1991, 49: 355-View ArticleGoogle Scholar
- Aller RC, Yingst JY, Ullman WJ: J Mar Res. 1983, 41: 571-View ArticleGoogle Scholar
- Pearson CE: J Math Phys. 1968, 47: 134-View ArticleGoogle Scholar
- Kristensen E, Hansen K: Biogeochemistry,. 1999, 45: 147-Google Scholar
- Miron G, Kristensen E: Mar Ecol Progr Ser. 1993, 101: 147-View ArticleGoogle Scholar
- Nithart M, Alliot E, Salen-Picard C: J Mar Biol Assoc UK. 1999, 79: 1029-10.1017/S0025315499001265.View ArticleGoogle Scholar
- Hansen K, Kristensen E: Estuarine Coastal Shelf Sci. 1997, 45: 613-10.1006/ecss.1996.0229.View ArticleGoogle Scholar
- Millero FJ: Limnol Oceanogr. 1986, 31: 839-View ArticleGoogle Scholar
- Millero FJ: Geochim Cosmochim Acta. 1995, 59: 661-10.1016/0016-7037(94)00354-O.View ArticleGoogle Scholar
- Davey JT: J Exp Mar Biol Ecol. 1994, 179: 115-10.1016/0022-0981(94)90020-5.View ArticleGoogle Scholar
- Rowe GT: Deep-sea Biology. Edited by: Rowe GT. 1983, John Wiley & Sons, New York, 97-Google Scholar
- Rosso L, Lobry JR, Bajard S, Flandrois JP: Appl Environ Microbiol. 1995, 61: 610-Google Scholar
- Heilskov AC, Holmer M: ICES J Mar Sci. 2001, 58: 427-10.1006/jmsc.2000.1026.View ArticleGoogle Scholar
- Aller RC, Aller JY: J Mar Res. 1998, 56: 905-10.1357/002224098321667413.View ArticleGoogle Scholar