A Model Study of the Effects of Sulfide-Oxidizing Bacteria

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Helsingin yliopiston digitaalinen arkisto

A model study of the effects of sulfide-oxidizing bacteria (Beggiatoa spp.) on phosphorus retention processes in hypoxic sediments: Implications for phosphorus management in the Baltic Sea

Sepehr Shakeri Yekta* and Lars Rahm

Department of Thematic Studies — Water and Environmental Studies, Linköping University, SE-581 83 Linköping, Sweden (corresponding author’s e-mail: [email protected])

Received 14 June 2010, accepted 23 Nov. 2010 (Editor in charge of this article: Heikki Seppä)

Yekta, S. S. & Rahm, L. 2011: A model study of the effects of sulfide-oxidizing bacteria (Beggiatoa spp.) on phosphorus retention processes in hypoxic sediments: Implications for phosphorus management in the Baltic Sea. Boreal Env. Res. 16: 167–184.

Ongoing eutrophication increases phosphorus storage in surficial sediments of the Baltic Sea which can then be released during hypoxic/anoxic events. Such sediments are suitable habitats for sulfide-oxidizing bacteria,Beggiatoa spp. The objective of this paper is to investigate the effects of these bacteria on the P retention processes in hypoxic sediments using a diagenetic model. This model simulates interactions of the processes controlling P mobility in the sediments with redox reactions from the Beggiatoa metabolism. Modeling results demonstrate that P retention capability is limited when dissolved iron is mineralized as iron sulfides in the sediments. In this regard, sulfide consumption by Beggiatoa spp. potentially decreases the rate of iron sulfide formation and consequently increases the P retention capability in local-scale sediment.

Introduction during the middle of the 1990s, mild windy winters led to an increased mixing of the water Surficial sediments of the Baltic Sea contain column, less hypoxic/anoxic bottoms, and a amounts of phosphorus (P) comparable to many drastic drop in P concentration in the water mass years of riverine loads (Conley et al. 2002). (Gustafsson and Stigebrandt 2007). This inspired This P source can be released during hypoxic/ the development of new remedy processes by anoxic events and in a longer perspective may improving the oxygen conditions in the bottom fuel nitrogen (N)-fixing cyanobacteria blooms layers and thereby increasing P retention in the and render remedy actions based on reduction sediments based on increased concentrations of of N loads less successful (Cloern 2001). In Fe(III) oxides (Mortimer 1941). Knowledge of fact, the portion of hypoxic/anoxic sediments the processes at the sediment water interface increased not only relatively but also quantita- regulating the hypoxic/anoxic sediments is cru- tively during the recent decades in the Baltic cial in these attempts. Sea. This increase seems to be a global phe- Hypoxic/anoxic bottoms eliminate higher nomenon (Diaz and Rosenberg 2008). However, orders of life but make the new sediment sur- 168 Yekta & Rahm • Boreal Env. Res. vol. 16 face a suitable habitat for sulfide-oxidizing Dissolution of iron oxide sheets to Fe2+ in the bacteria, Beggiatoa spp. (Rosenberg and Diaz top layers of the sediments caused by shifts from 1993). Beggiatoa cells build aggregates in sedi- oxic to hypoxic/anoxic conditions in bottom ments rich in organic matter with a gradient water, changes the sorption capability of the sed- of sulfide across the sediment water interface iments and leads to mobilization of iron-bound (Mußmann et al. 2003). These bacteria are che- phosphate which can be transferred to overlying moautotrophs, gaining metabolic energy by water columns (Conley et al. 2002). oxidizing hydrogen sulfide (Jørgensen 1982). To understand the effects of Beggiatoa mats Widespread Beggiatoa mats have been observed on P retention processes, we developed a dia- in the Stockholm Archipelago (Rosenberg and genetic model for reactive diffusive transport Diaz 1993), in anoxic sediments of the Gulf of of substances to simulate major biogeochemical Finland (Vallius 2006), and in huge areas of the processes in the sediments like those in the Baltic eastern Gotland basin (P. Hall, Gothenburg Uni- Sea with both Beggiatoa-free and Beggiatoa- verity, pers. comm.). The Danish Limfjorden, the covered sediment surfaces. The Beggiatoa-free German Wadden Sea and Eckenförde Bay are sediment model couples biotic processes such other examples where large populations of these as organic matter degradation pathways with bacteria have been observed (Mußmann et al. inorganic processes including iron oxide precipi- 2003, Preisler et al. 2007). tation/dissolution, iron sulfide precipitation, and Beggiatoa spp. is well-known for protecting P adsorption/desorption. Organic matter degra- overlying water from toxic sulfide diffused from dation pathways are the main contributors to the sulfidic layers of the sediments. Beggiatoa redox disequilibrium in the sediments (Tromp metabolism including their physiology and enzy- et al. 1995), while inorganic processes are the mology have been studied in detail (e.g. Jør- major chemical pathways controlling mobility gensen 1977, Hagen and Nelson 1997, Mußmann of P in marine sediments (Jensen et al. 1995, et al. 2003, Sayama et al. 2005, Hinck et al. Carman and Rahm 1997, Gunnars et al. 2002, 2007, Kamp et al. 2008). Furthermore, the role of Conley et al. 2002, Lehtoranta and Pitkänen these bacteria in the cycling of elements has been 2003, Lehtoranta et al. 2008). In the next step, the topic of numerous studies. As an example, we added microbially mediated redox reactions Fenchel and Bernard (1995) investigated major from the Beggiatoa metabolism to the model in microbial processes such as nitrification/denitrifi- order to simulate their interactions with sedi- cation and cycling of N in a Beggiatoa inhabited mentary processes. The development and appli- sediment. Preisler et al. (2007) studied sulfide cation of diagenetic models for the investigation oxidization by Beggiatoa mats and its interac- of sedimentary processes can be found in e.g. tions with manganese and iron in local scale sedi- Berner (1964), Boudreau (1997), and Boudreau ment. Most of the previous studies emphasized and Jørgensen (2001). the importance of these bacteria in the cycling The current model simulates the key P reten- of elements in marine environments, but to our tion processes in natural sediments based on knowledge, none of the earlier studies investi- observations and field measurements of water gated the impact of Beggiatoa spp. on P retention and sediments of the Baltic Sea. Through mode- processes in their growth medium. ling, we do not intend to reproduce the observed Phosphorus (P) retention processes in the data but to provide a realistic numerical rep- sediments of the Baltic Sea have been amply resentation of the sedimentary processes. The researched. For instance, Brügmann et al. goal of the study is to investigate our conceptual (1992), Sternbeck and Sohlenius (1997), Hille understanding of the impact of Beggiatoa mats et al. (2005), and Leipe et al. (2008) have all on the P retention processes in sediments by contributed to knowledge in this area. Previous comparing two conventional diagenetic models studies showed that mobility of P in the sedi- simulating typical Baltic Sea hypoxic sediments ment–porewater medium is chiefly regulated by which differ only in the presence or absence of adsorption of phosphate on iron oxyhydroxides. Beggiatoa mat. Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 169

Beggiatoa filaments at Material and methods Mixed water the sediment surface

We developed a 1D diagenetic model to simulate Metabolic activities DBL the vertical distribution of solute and solid com- – 2+ 2+ z Porewater: O2, NO3 , Mn , Fe , 2– + 3– pounds in local scale sediment. In doing this, SO4 , NH4 , H2S, PO4 we treated sedimentary solute and solid species Organic and inorganic redox (Fig. 1) as dependent variables and considered reactions, adsorption/desorption, their mass conservation equations in an arbitrary precipitation/dissolution sediment system. We included chemical reaction Sediment Organic and mineral particles, pathways in conservation equations as sources organic carbon, Fe(OH)3, MnO2, FeS or sinks to simulate dynamic interactions of dis- Fig. 1. Model conceptualization and incorporated bio- solved and solid compounds. Boudreau and Jør- chemical processes in the modeled sediment column. z gensen (2001) discussed derivation of mass con- represents sediment depth. servation equations for sedimentary substances and biogeochemical processes in detail. 1997) in order to compare simulated and measured concentration profiles and adjusted some Model conceptualization model parameters to approach the measured profiles. This provided us with a numerical tool for Model conceptualization including incorporated further investigations of the P retention processes biochemical reactions is illustrated in Fig. 1. The as well as the effects of Beggiatoa mats on them. model simulates an arbitrary vertical column of a sediment–porewater system with a given geometry. The present model simulates the dis- Physical model properties tribution of dissolved and solid compounds from the mixed water body downwards across the The physical properties of the model are listed diffusive boundary layer (DBL) and into the in Table 1. We fixed the model geometry as a sediment-porewater medium. The model was vertical sediment column with a depth of 20 cm. allowed to spin up during a simulation period The system was divided into layers of 0.1 cm of four years which is long enough to produce thickness. The thin layers were required to cap- quasi-steady concentration profiles by time. We ture steep gradients below the surface sediment. also applied field data from porewater analysis A time step of six hours was chosen. Decreas- of three hypoxic sediments (Carman and Rahm ing the time step to less than six hours did not

Table 1. Physical parameters of the model. Explanations: µ is the dynamic viscosity, Vb is molar volume of the nonelectrolyte, m0 and m1 are experimentally measured linear regression coefficients for calculation of diffusion coefficients of cations and anions against temperature (τ, °C) based on the Nernst-Einstein expression, T is the temperature in kelvins (K) . Values for these parameters are found in Boudreau (1997).

Parameter Value or formula (unit) Source

Sediment depth 20 (cm) Length of iterations 0.1 (cm) Length of time steps 6 (hr) Simulation period 4 (yr) Porosity φ = 0.8 Carman and Rahm (1997) Sedimentation rate w = 0.1 (cm yr –1) Hille et al. (2006) Tortuosity θ2 = 1 – lnφ Tsai and Strieder (1986) 0.69 2 –1 Biodiffusion coefficient DB = 15.7w (cm yr ) Boudreau (1997) ¥ –9 –0.6 2 –1 Diffusivity of molecules in water Dmolecular = 4.72 10 [T( µVb) ] (cm yr ) hayduk and Laudie (1974) ¥ –6 2 –1 Diffusivity of ions in water Dionic = (m0 + m1τ) 10 (cm yr ) Boudreau (1997) 170 Yekta & Rahm • Boreal Env. Res. vol. 16 change the simulated profiles of the compounds. beneath the sediment–water interface. Organic Hille et al. (2006) estimated the mean linear matter degradation with multiple electron accep- sedimentation rate of roughly 0.1 (cm yr–1) for tors is the most important microbial process con- the Gotland basin. We calculated diffusion coef- trolling redox conditions in the sediments (Tromp ficients of porewater species according to the et al. 1995). In the current model, we considered formula derived for the diffusion of correspond- plankton decomposition as the ultimate energy ing molecules or ions in the water (Table 1) supply for the microbial community (Redfield and modified their values for a porous medium 1958). In general, over 90% of deposited organic by including the parameter of tortuosity. This matter is oxidized in the upper layers of the parameter considers the longer distances that sediments (15–20 cm), and only a small, inert a molecule has to travel in a porous medium fraction is permanently buried (Jørgensen 1982, compared to water. We used the average values Boudreau 1997). The rates of different organic of bottom water composition from the three carbon oxidization pathways primarily depend hypoxic stations studied by Carman and Rahm on the types of available organic matter in terms (1997) in order to estimate both the upper bound- of reactability. Reactive organic matter is easily ary conditions and the initial values of the model degraded while the refractory fraction is resistant variables (Table 2). Organic and inorganic reac- to microbial assimilation and consequently, its tions and the rationale for choosing them are degradation rate is lower than the reactive frac- described in the following subsections. tion (Tromp et al. 1995). We applied first order kinetics for the oxidization rate of organic matter (Berner 1964, Canfield et al. 1993, Tromp et al. Organic matter degradation processes 1995, Van Cappellen and Wang 1996):

Organic detritus from the water column is the RO = kCorgC (1) energy source for diverse microorganisms living

where CorgC is the concentration of organic carbon and k is the generic rate constant for Table 2. Upper boundary conditions and initial values of reactive (kreactive) and refractory (krefractory) organic sediment and porewater compounds. carbon calculated according to formula reported Dissolved compound aBottom water by (Tromp et al. 1995): concentration (µM) 0.62 kreactive = 2.97w (2) O2 91 – NO3 10.27 1.94 2+ b Mn 0 krefractory = 0.05w (3) Fe2+ 0b 2– SO4 8500 where w is the linear sedimentation rate. In NH + 0.38 4 marine sediments, electron acceptors with higher H2S 0 3– PO4 2.41 Gibbs free energy suppress the reactions of oxidants with lower Gibbs free energy (Van Cap- c Solid compound Deposition rate pellen and Wang 1996). In oxic layers, oxygen, (µmol cm–2 yr–1) as the most powerful oxidant, is responsible for Organic C 2500 organic matter degradation through the respira-

Fe(OH)3 60 tion of aerobes. Depletion of oxygen favors other MnO 40 2 electron acceptors and shifts the organic carbon FeS 0 oxidization pathway from aerobic respiration to a Average values of bottom water concentrations from denitrification and subsequently to Mn and Fe Carman and Rahm (1997) oxide reduction, sulfate reduction and methano- b From Canfield et al. (1993). genesis. These processes are listed in (Table 3). c Deposition rates were adjusted for calibration purposes and initial values of the solid compounds are Sequential inhibition of organic carbon oxi- assumed zero. dization pathways by more powerful electron Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 171

Table 3. Chemical expressions of organic matter degradation pathways (Redfield 1958, Tromp et al. 1995, Van Cappellen and Wang 1996).

Pathway Stoichiometry of reaction

– → – R1: Aerobic respiration G + (x + 2y)O2 + (y + 2z)HCO3 (x + y + 2z)CO2 + (x + 2y + 2z)H2O + yNO3 2– + zHPO4 – → – R2: Denitrification 5G + (4x + 3y)NO3 (x – 3y + 10z)CO2 + (3x + 6y + 10z)H2O + (4x + 3y – 10z)HCO3 2– + (2x + 4y)N2 + 5zHPO4 → 2+ – R3: Mn oxide reduction G + 2xMnO2 + (3x + y – 2z)CO2 + (x + y – 2z)H2O 2xMn + (4x + y – 2z)HCO3 + 2– + yNH4 + zHPO4 → 2+ – + R4: Fe oxide reduction G + 4xFe(OH)3 + (7x + y – 2z)CO2 4xFe + (8x + y – 2z)HCO3 + yNH4 2– + (3x – y + 2z)H2O + zHPO4 2– → – + R5: Sulfate reduction G + (x/2)SO4 + (y – 2z)CO2 + (y – 2z)H2O (x/2)H2S + (x + y – 2z)HCO3 + yNH4 2– + zHPO4 → – + 2– R6: Methane production G + (y – 2z)H2O (x/2)CH4 + [(x – 2y – 4z)/2]CO2 + (y – 2z)HCO3 + yNH4 + zHPO4

G: Plankton composition, (CH2O)x(NH3)y(H3PO4)z; x = 106, y = 16, z = 1.

acceptors can be included into diagenetic models where RO is the first order oxidization rate of by the introduction of “inhibition constants” to organic carbon (Eq. 1), Ci is the concentra- the rate equations of organic matter degradation tion of the limiting electron acceptor i (i.e. O2, – 2– (e.g. Boudreau and Westrich 1984, Tromp et NO3 , MnO2, Fe(OH)3, SO4 ), and KSi is the al. 1995). To avoid overparameterization of the corresponding saturation constant. We treated model due to the application of this approach, we the saturation constants as model parameters and implemented a decision algorithm in the com- used them for calibration purposes (Table 4). puter program to shift the oxidization pathways while the pool of an electron acceptor with higher oxidization power is depleted in the sediment Inorganic processes medium (Van Cappellen and Wang 1996). We assumed that organic matter decomposition does The model couples organic carbon oxidization to not occur via two simultaneous processes except inorganic reactions. In particular, we incorporated for nitrate reduction which occurs under both the most important P retention processes in the aerobic and anaerobic conditions (Robertson and model (Table 5). In the sediments, iron oxides, Kuenen 1984). The limiting effects of the con- as the main P-bearing minerals, are microbi- centration of porewater electron acceptors on ally (Table 3: R4) and chemically (Table 5: corresponding organic degradation pathways are incorporated by including the Monod kinetics Table 5. Inorganic reaction network in the modeled term into the organic carbon oxidization rate: sediment (Van Cappellen and Wang 1996, Boudreau 1997). (4) Stoichiometry of reaction

→ 2+ – R7: h2S + 2Fe(OH)3 + 4CO2 2Fe + 4HCO3 0 Table 4. Modeled values of saturation constants for dif- + S + 2H2S → 2+ – 0 ferent organic matter degradation pathways. R8: h2S + MnO2 + 2CO2 Mn + 2HCO3 + S – → 2– R9: h2S + 2O2 + 2HCO3 SO4 + 2CO2 + 2H2O 2+ – → Reaction number saturation constant Unit R10: Fe + 1/4O2 + 2HCO3 + 1/2H2O Fe(OH)3

+ 2CO2 2+ – → R1 µM R11: mn + 1/2O2 + 2HCO3 MnO2 + 2CO2 + H2O 2+ – → R2 µM R12: 2Fe + MnO2 + 2HCO3 + 2H2O 2Fe(OH)3 –1 2+ R3 µmol g + Mn + 2CO2 –1 + – → – R4 µmol g R13: nh4 + 2O2 + 2HCO3 NO3 + 2CO2 + 3H2O 2+ – ↔ R5 mM R14: Fe + 2HCO3 + H2S FeS + 2CO2 + 2H2O 172 Yekta & Rahm • Boreal Env. Res. vol. 16

R7) reduced to ferrous ions by iron(III)-reducer P adsorption bacteria and sulfide respectively. In oxic parts, ferrous ions are mineralized by oxygen (Table 5: Adsorption of various porewater phosphate (PP) R10) and freshly precipitated iron oxides adsorb species on the sorption sites of iron oxide min- porewater phosphate (Sobolev and Roden 2001). erals and formation of iron-bound phosphate Furthermore, Mn oxides increase the efficiency (IBP) species is the main reaction controlling of the iron oxide precipitation process through the mobility of P in the Baltic Sea sediments their reductive dissolution by ferrous ions (Table (Jensen et al. 1995, Conley et al. 2002, Gun- 5: R12) (Myers and Nealson 1988). The pre- nars et al. 2002, Lehtoranta et al. 2008). Several cipitation/dissolution of iron sulfide minerals is studies tried to identify various PP and IBP spe- described in (Table 5: R14). The process depends cies in this process (e.g. Tejedor-Tejedor and on the surface area and roughness of the mineral Anderson 1990, Persson et al. 1996, and Gao phase as well as the chemical effects of the pore- and Mucci 2003). According to these studies, water composition which are incorporated in the PP includes compounds from the dissociation – 2– 3– rate constants of the reactions (Rickard 1995). of phosphoric acid (H2PO4 , HPO4 , and PO4 ) We applied bimolecular kinetics to correlate as well as a variety of ion-pairs formed between the rates of homogeneous and heterogeneous these substances and porewater cations (e.g. + 2– inorganic reactions with the porewater concen- MgHPO4, CaH2PO4 , and NaPO4 ). IBP includes tration of redox couples (Van Cappellen and monodentate and bidentate surface complexes Wang 1996): formed through interactions of PP species with hydroxyl groups of ferric oxide minerals [e.g. – ROx.Rd = kOx,RdCOxCRd (5) FeH2PO4 and (FeO)2PO4 ]. In a multicomponent electrolyte system like where kOx,Rd is the rate constant and C is the con- the sediment-water medium, porewater proper- centration of denoted oxidant (Ox) and reductant ties such as pH, ionic strength, and the presence (Rd). Rate constants from Van Cappellen and of competitive adsorbents influence the distribu- Wang (1996) were used as reference values and tion of phosphate between sediment mineral and we adjusted the rate constants so the modeled porewater phases. Furthermore, types of iron profiles of the sedimentary substances would oxide particulates (e.g. ferrihydrite, goethite and correspond with the observed trends (Table 6). hematite), their availability, structure, and sorp- It is noted that the calibrated kinetic parameters tion affinity are influential on the adsorption/des- reflect the model’s uncertainties and simplify- orption of phosphate (Geelhoed et al. 1997, Gao ing assumptions hence they are not applicable and Mucci 2001, Spiteri et al. 2008). To simplify in other studies as information for individual this complexity, we applied the main features of chemical processes. the adsorption equilibrium herein. It is assumed that iron oxide minerals have an equal potential for P sorption and a homogeneous surface with Table 6. Rate constants of inorganic reactions. hydroxyl groups being the only active surface functional group. The model defines the param- Reaction adjusted value value from number (Van Cappellen eters of PP and IBP as representatives for all and Wang 1996) porewater and iron-bound P species respectively, (µM–1 yr–1) irrespective of their particular type:

R7 ≤ 0.1 PP = H PO – + HPO 2– + PO 3– + ∑ion-pairs (6) R8 ≤ 0.01 2 4 4 4 R9 ≥ 0.16 R10 140 IBP = ∑ Fe  PO4 (7) R11 0.8–20 R12 ≤ 1000 where the symbol  represents sorption sites on R13 10 the surface of iron oxide minerals. The model R14 no data assumes equilibrium adsorption/desorption reac- Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 173 tions which implies the rate of these reactions dissolved and iron-bound forms: are fast relative to reactive diffusive transport of substances in the sediment (Spiteri et al. 2008): CTP = CPP + CIBP (11)

R15: PP +  FeOH ↔ IBP The conservation equation for this variable is derived by summing mass conservation equa- We used the equilibrium isotherm to describe tions of PP and IBP as follows: the partitioning of porewater and iron-bound fractions of phosphate:

CIBP = KPCPPCFeOH (8) (12) where C is the concentration of the denoted solute and solid substances, and KP is termed as the overall equilibrium constant and treated where t is time, z vertical coordinate (depth), C as a model parameter (with the calibrated value concentration of denoted species, w sedimenta- 2.5 of 10 ). Recall that the present P adsorption tion rate, θ tortuosity, ∑RP the sources and sinks model neither considers the types of surface of TP in form of chemical reactions, and DB the complexes and dissolved species nor their indi- biodiffusion coefficient see( Table 1). DPP is the vidual adsorption equilibrium. Furthermore, the diffusion coefficient of dissolved phosphate spe- introduced reaction (R15) represents the overall cies including ion pairs calculated according to adsorption equilibrium instead of specifying this the formula reported by Krom and Berner (1980): process for individual solid and solute phosphate species. KP incorporates the effects of pH, ionic DPP = 112(1 + 0.048τ) (13) strength and surface charging on the availability of free surface sites for P adsorption. According where τ is ambient temperature of sediment– to the presented adsorption scheme, fractions of water system in °C. PP and IBP and their correlation with available sorption sites (CFeOH) are determined as follow: Beggiatoa metabolism (9) The current model couples microbially mediated redox reactions from the Beggiatoa metabo- (10) lism with sedimentary multicomponent reactions (Fig. 1). According to early studies of Beggia-

We termed λPP and λIBP as the distribution toa spp., these microorganisms are characterized coefficients for porewater and iron-bound phos- by their sulfide oxidization capability, internal phate species respectively, which quantify PP sulfur vacuoles and gliding motility. They obtain and IBP fractions in any given depth of the sedi- metabolic energy by the aerobic oxidization of ment according to available sorption sites. sulfide to internal sulfur and/or sulfate. They use Since we intended to investigate the major P their gliding motility to trace the oxic/anoxic retention processes in the sediments, less influ- interface as a way for balancing fluxes of sulfide ential processes such as adsorption of phosphate and oxygen across their micrometer scale habitat on other sedimentary particulates like Mn oxides (Jørgensen 1982, Jørgensen and Revsbech 1983, and the formation of authigenic P-containing Nelson et al. 1986). minerals (Ruttenberg and Berner 1993) were not More recently, Beggiatoa enzyme analyses considered. The latter was negligible in analyzed demonstrated that the biomass of these micro- sediments of the Baltic Sea by Carman and organisms contains respiratory nitrate reductase Rahm (1997). Hence, total phosphate (TP) in the which enables them to use nitrate as an electron modeled sediment system is the summation of acceptor for sulfide oxidization (McHatton et al. 174 Yekta & Rahm • Boreal Env. Res. vol. 16

1996). Beggiatoa filaments are able to internally purposes. Therefore, net biomass accumulation accumulate nitrate in levels up to four orders of is zero and the volume of the cells remains con- magnitude higher than the nitrate concentration stant during the simulation period. of their surrounding porewater (Mußmann et al. Although Beggiatoa filaments respire oxygen, 2003) and to use intracellularly stored nitrate their microaerophilic characteristic forces them for sulfide oxidization. Beggiatoa nitrate storing to respond to high oxygen concentrations in the ability together with their gliding motility ena- water (Møller et al. 1985). To cope with this bles them to trace and take up sulfide in anoxic inconsistency, these microorganisms adjust their parts of the sediments and expand their living vertical position in order to control the oxygen environment from the micrometer to the centim- flux at the cell membrane. In highly oxic condi- eter scale (Kamp et al. 2006). tions, the filaments avoid high oxygen concentra- Beggiatoa filaments integrate gliding motil- tions by converging below the DBL and forming ity with nitrate and sulfur storage to efficiently a dense and smooth bacterial mat. When oxygen scavenge substrates from the environment. When is limited under hypoxic conditions, the Beggia- oxygen and nitrate are available, Beggiatoa fila- toa community extends its filaments above the ments perform sulfide oxidization with two spa- sediment surface to capture more oxygen (Jør- tially separated pathways (Table 7). Beggiatoa gensen 1982, Møller et al. 1985). Under these filaments store nitrate intracellularly (Table 7: conditions, the rougher surface sediment formed R16), in vacuoles or possibly cytoplasm (Kamp due to extended Beggiatoa filaments hinders flow et al. 2006) and transport it downwards to anoxic in the viscous boundary layer and increases both sediment layers via their gliding mechanism. the thickness of DBL (Jørgensen and Revsbech Transported nitrates are used for anaerobic 1983) and the diffusion distance for substances sulfide oxidization to elemental sulfur (Table 7: from the water column to the sediment. R17) which will then be stored in vacuoles for Møller et al. (1985) measured variations of aerobic oxidization to sulfate in oxic parts of the 0.5 to 1.1 mm in DBL thickness over a Beggia- bacterial mat (Table 7: R18) (McHatton et al. toa mat at high and low oxygen fluxes respec- 1996, Sayama et al. 2005, Kamp et al. 2006). tively under laboratory conditions. We used an The model implements the chemical mech- average value of 0.8 mm in the model to calcu- anism proposed by Sayama et al. (2005) to late fluxes of different solutes across the DBL. describe the Beggiatoa metabolism (Table 7). Diffused oxygen and nitrate from the overlying This mechanism characterizes the “energy main- water are exposed to a chemically active layer tenance phase” of Beggiatoa filaments where the of Beggiatoa filaments while passing across the sulfide is completely oxidized to sulfate and the DBL (Fig. 2). We modified the upper boundary generated energy is sufficient for cellular activity conditions for oxygen and nitrate by incorporat- (Sayama et al. 2005). In the model, we assumed ing bacterial uptake rates into the mass conserva- that the rate of biomass increase due to bacterial tion equations used for the derivation of sedi- growth is equal to its decrease due to cell death ment surface boundary conditions. and endogenous metabolism including processes As a consequence of the Beggiatoa response leading to a decrease in cell mass such as the oxi- to environmental variations, these microor- dization of internal sulfur and nitrate for energy ganisms impose spatially fluctuating chemical changes to the sediment environment. These changes result in a spatial separation of oxic Table 7. Chemical mechanism proposed by Sayama et and sulfidic zones with 1–3 cm thickness by al. (2005) to describe the Beggiatoa metabolism. nitrate-storing Beggiatoa (Mußmann et al. 2003, Stoichiometry of reaction Sayama et al. 2005, Kamp et al. 2006) or an overlap of oxygen and sulfide concentration pro- – → – R16: (NO3 )dissolved (NO3 )intercellular files in micrometer scales when oxygen is the – + → 0 + R17: 4H2S + (NO3 )intercellular + 2H 4(S )intercellular+ NH4 only available oxidant (Nelson et al. 1986). The + 3H O 2 model assumes a mature Beggiatoa community R18: 4(S0) + 6O + 4H O → 4SO 2– + 8H+ intercellular 2 2 4 (i.e. the growth phase is passed) which uses both Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 175

– oxygen and nitrate for its metabolism and is O2 and NO3 Water able to store nitrate and extends its filaments to Elevated chemical DBL deeper sediment layers for capturing sulfide. Our interface modeled Beggiatoa culture inhabits the sediment z = 0 Sediment surface layers down to the depth of 2 cm. We Beggiatoa further assumed that the Beggiatoa position is filaments steadfast during the simulation period. Fig. 2. Conceptualization of the elevated oxygen and Due to a spatial separation of the microbial nitrate chemical interface due to upward migration of matrix from the overlying turbulent water by the Beggiatoa filaments under hypoxic conditions. DBL, oxygen and nitrate transport from water to the Beggiatoa mat takes place by diffusion (Jør- gensen and Revsbech 1983). Diffusion of sub- researchers (Table 8). The model assumes that strates to the Beggiatoa surface membrane is a both oxygen and nitrate fluxes in the top layers of complex multidirectional transport of molecules. the sediment and the sulfide production rate are We simplified this complexity by incorporating the major environmental factors regulating the uptake rates for the whole bacterial population Beggiatoa metabolism. Hence, Jenv.j is equivalent into the reaction network of the sediment. We to the fluxes of these substances in any given used the Monod-type kinetics to represent bac- depth of the sediment occupied by Beggiatoa terial consumption rates of sulfide, oxygen and filaments corresponding to the maximum flux nitrate in the following model: rates of substrates available for bacterial consumption. Taking into account the internal stor- (14) age of nitrate for sulfide oxidization and sulfur for oxygen uptake, we assumed that the supply where Cj is the extracellular concentration of of these compounds does not limit the metabolic substrate j (i.e. sulfide, oxygen, and nitrate), and activities. Furthermore, the model treats Beggia-

Jenv.j is flux rate of substrate j across the Beggia- toa biomass as a non-limiting factor. toa habitat. The Monod type function incorpo- It is noteworthy that the Beggiatoa effects on rates the limiting effects of ambient concentra- the biogeochemistry of the sediments are highly tions on uptake rates. The adjustable parameter dependent upon their density (Preisler et al. 2007). of Kaf.j incorporates the limitations regarding In the current model, we did not specify a den- binding affinity of the substrate j (Shuler and sity for the Beggiatoa community and assumed

Kargi 2002). We assigned values of Kaf.j for that both the distribution of Beggiatoa biomass different substrates in a way that the model and their chemical effects are uniform in their produces overall substrate uptake rates compara- inhabited sediment and they are able to take up ble to experimentally measured values by other substrates with maximum rates equal to the fluxes

Table 8. Comparison of measured uptake rates of enriched Beggiatoa filaments by (Kampet al. 2006) and modeled values.

Substrate Uptake rate (µM day–1) comment

Sulfide 514 After two days of experiment Sulfide 186 After four days of experiment Sulfide 194 Model Oxygen 155 After two days of experiment Oxygen 332 After four days of experiment Oxygen 230 Model Nitrate 257 Decrease in nitrate concentration after two days of experiment from initial value of 600 µM to 86 µM Nitrate 16 Decrease in nitrate concentration from 86 µM to 54 µM between day two and day four of experiment Nitrate 26 Model 176 Yekta & Rahm • Boreal Env. Res. vol. 16

Finland

Sweden 60°N BY31s Russi a Estonia

Depth (m) 58°N 0 Latvia BY9

–60

56°N k Lithuania –120

BCS III-10 Denmar –180 Fig. 3. Schematic map showing the location of sampling stations in the –240 Baltic Sea by Carman and 54°N Poland Germany Rahm (1997). Bathymetric data is from Seifert and 11°E 15°E 19°E 23°E 27°E 31°E Kaiser (1995). of substances across their habitat. For simplicity, bottom-water oxygen concentrations were 82.5, we also assumed that the physical effects of these 109.3, and 81.2 µM, respectively. BCS III-10 is bacteria on the transport of solute and solid com- located at the eastern end of the Stolpe Furrow. pounds are negligible in the modeled sediment. The topmost layer of the sediment was light Furthermore, seasonal variations in the Beggia- brown in color and the lower layers were indis- toa density, the regulatory effects of sediment tinctly laminated and contained “black striae”. structure (Jørgensen 1977), and the role of other BY9 is located at the southern end of the eastern benthic microorganisms are examples of influen- Gotland basin. The upper 4 cm of the sediment tial parameters on Beggiatoa metabolism which was light brown and the pronounced lamination are not captured by the model. In this regard, the was observed in the sediment between the depths modeling experiment should not be considered a of 4 and 10 cm. BY31s is located at the edge of method for simulating the exact Beggiatoa behav- the Landsort Deep and the sediment from this ior in natural sediments, but rather as a multi- station had brownish top layers and contained variable approach for investigating their potential laminated layers with “black striae” down to the influences on sedimentary processes. depth of 23 cm. The light brown color of the top layers of the sediments and the observed “black striae” suggest the presence of iron oxide and Results and discussion iron sulfide minerals, respectively (Carman and Rahm 1997). Site description

Carman and Rahm (1997) analyzed surficial The case of Beggiatoa-free sediment porewater composition from seven deep basins in the Baltic Sea. We considered three of the sites Modeled concentration profiles of compounds — BCS III-10, BY9, and BY31s — to be repre- as functions of sediment depth are presented in sentative of hypoxic sediments (Fig. 3). These Fig. 4 together with measured porewater com- sites have water depths of 91, 123, and 125 position and organic matter contents from sedi- meters, respectively. At the time of sampling, the ments of BCS III-10, BY9, and BY31s. We arbi- Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 177

0 0 0 a b c 2 2 2

4 4 4

6 6 6 BY31s

Sediment depth (cm) BY9 8 8 8 BCS III-10 Model 10 10 0 10 20 30 40 50 0 50 100 100 200 400 600 800 Fe(II) (µM) Mn(II) (µM) Sulfide (µM) 0 0 0 d e f 2 2 2

4 4 4 Fig. 4. Modeled and observed concentration profiles of sedimentary 6 6 6 substances. a–g: compar- Sediment depth (cm) ison of the model results 8 8 8 after four years simulation period (dashed lines) with 10 10 10 three hypoxic sediment 0 200 400 600 0 2000 4000 6000 8000 0 100 200 300 400 Ammonium (µM) Sulfate (µM) PP (µM) profiles, BCS III-10, BY 0 0 0 9, and BY 31s (Carman and Rahm, 1997). h and g h i i: model results showing 1 1 1 sequential depletion of oxygen, nitrate, Mn oxide, 2 2 2 and Fe oxide through corresponding organic matter 3 3 3 degradation pathways. Sediment depth (cm) Please note the different 4 4 4 scales of the profiles of Oxygen Fe oxide organic carbon, oxygen, Nitrate Mn oxide 5 5 nitrate, Fe oxide and Mn 0 2000 4000 6000 0 1 2 3 4 5 50 20 40 60 oxide. Organic carbon (µmol g−1) Concentration (µM) Concentration (µmol g−1) trarily chose a simulation period of four years to Simulated profiles of electron acceptors for calculate quasi-steady profiles of sedimentary organic carbon oxidization demonstrated that compounds. It was not our intention to repro- oxygen and nitrate (oxidants with the highest duce the observed data, but rather to provide a Gibb’s free energy) were completely depleted realistic representation of the sediment and pore- in the upper 1 cm of the modeled sediment. A water profiles based on the dominant processes suboxic zone developed within the depth of in the sediments. Simulated concentration pro- 1 to 2 cm, where Mn and Fe oxides were files (dashed lines) reasonably correspond to the terminal electron acceptors. Microbial reduc- observed trends of porewater Fe2+ (Fig. 4a), Mn2+ tion of these metal oxides resulted in a sharp (Fig. 4b), sulfide (Fig. 4c), ammonium (Fig. increase in Fe2+ and Mn2+ concentrations, reach- 4d), sulfate (Fig. 4e), PP (Fig. 4f) and the top ing their maximum levels in suboxic zones (Fig. layers’ organic content (Fig. 4g) together with 4a and b, respectively). Subsequent to iron oxide the sequential depletion of electron acceptors for depletion, sulfate became the terminal electron organic carbon oxidization and the formation of acceptor and the organic matter degradation suboxic and anoxic zones (Fig. 4h and i). pathway shifted to sulfate reduction where the 178 Yekta & Rahm • Boreal Env. Res. vol. 16

0 0 a b 1 1

2 2

3 3

4 4

5 5

6 6

Sediment depth (cm) 7 7 Fig. 5. Modeled concen- 8 8 PP (µM) tration profiles of (a) PP, −1 Fe oxides and TP; and (b) 9 Fe oxide (µmol g ) 9 IBP TP (µM) PP distribution coefficients 10 10 of PP (λPP) and IBP (λIBP) 0 50 100 150 0 0.2 0.4 0.6 0.8 1 as functions of sediment Concentration Distribution coefficient depth (Eqs. 9 and 10). anoxic zone was developed. In this zone, below oxide contents of the sampled sediments to com- 3 cm, high rates of sulfate reduction resulted pare actual and calculated values. However, the in extreme sulfide gradients and upward fluxes availability of iron oxides was evident according (Fig. 4c). It is noteworthy that the simulated to the observed light brown color in the upper profiles are the results of coupled organic and layers of the sediments analyzed by Carman and inorganic processes and represent their overall Rahm (1997). The model simulated formation interactions. of iron sulfide with a depth integrated value of All pathways of organic matter degradation 33 µmol g–1 at the end of the simulation period. act as sources of phosphate in the sediments We assumed that deposition rate and initial con- (Table 3: R1–R6). Simulation results showed an centration of these minerals were zero. Hence, increase in TP contents (summation of IBP and the calculated concentration reflects iron sulfide PP) from the sediment–water interface down to minerals of authigenic origin. Although no quan- lower depths (Fig. 5a). In the upper 1 cm of the titative measurements are available regarding modeled sediment, the distribution coefficient of iron sulfide concentrations in the sites, the “black

IBP (λIBP) was equal to 1, indicating that in the striae” observed in the lower layers suggest the presence of iron oxide minerals, the TP content precipitation of these minerals in the hypoxic of the modeled sediment was in iron-bound sediments of the Baltic Sea. forms (Fig. 5a). Corresponding to the depletion The results from investigations of the P reten- of the pool of iron oxides in suboxic layers, λIBP tion processes in the sediments are presented as decreased to 0 and λPP increased to 1, represent- a simple conceptual model in Fig. 6. The figure ing the transfer of phosphate from iron-bound depicts major processes and compounds control- forms to dissolved forms (Fig. 5b). These results ling P mobilization in the sediments. Through imply that the transition from oxic (λIBP = 1) to oxidization of ferrous ions with oxygen and suboxic (0 < λIBP < 1) and anoxic (λIBP = 0) con- Mn oxides, sediment is enriched with Fe oxides ditions resulted in reductive dissolution of iron (pathway A) and porewater P contents are par- oxide minerals and mobilized phosphate in the tially or completely adsorbed on sorption sites modeled sediment environment. of these minerals depending on the extent of the The simulated concentration profile of iron iron oxide precipitation process (pathway C). oxides reflects the bioavailable portion of these However, dissolution of Fe oxides to iron ions minerals which are also active in the phosphate through chemical and microbial reduction (path- adsorption process. No data is available on iron way B) releases iron-bound P in the sediment Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 179

Transport processes Feoxide A: O2, MnO2 Adsorbed P Fig. 6. A conceptual model presenting major P reten- Organic matter degradation pathways tion processes in the sedi- D C ments. (a) Fe oxide pre- : Fe(III)-oxide cipitation by oxygen and B reducer bacteria, H2S Mn oxide, (b) Fe oxide Porewater P dissolution by Fe(III)- oxide reducer bacteria and sulfide, c ( ) P adsorption on Fe oxide miner- FeS Dissolved Fe als, (d) P desorption from x Fe oxide minerals, (e) Fe sulfide precipitation. E: H2S

(pathway D). Through the reaction of dissolved ered by a Beggiatoa mat with the characteris- iron with hydrogen sulfide (pathway E) as a tics described above. All model properties were subsequent reaction to the chemical dissolution kept untouched and the uptake rates (Eq. 14) of of iron oxide by sulfide (Rickard 1995), iron ions sulfide, oxygen, and nitrate as well as Beggiatoa are converted into iron sulfide minerals resulting production rates of ammonium and sulfate were in the fixation of iron molecules in non-reactive included in the sedimentary chemical network forms regarding P adsorption. according to reactions R16 to R18 (Table 7). To summarize, the conceptual model dem- The model was then left for the same simulation onstrates that the mobilization (pathway D) and period of four years as before to reach a quasi- immobilization of P (pathway C) in the sedi- steady state condition. The results are presented ments are mainly regulated by the extent of in Fig. 7, which includes model results from both the processes converting the iron supply of the Beggiatoa-free and Beggiatoa-inhabited sedi- sediment to reactive iron oxides or non-reactive ments and demonstrate the changes in concentra- iron sulfides, in terms of phosphate adsorption tion profiles when the Beggiatoa metabolism is (pathways A, B and E). Similarly, field studies added to the sedimentary chemical network. by other researchers demonstrated that the incor- It turns out that the net impact of Beggiatoa poration of iron ions into iron sulfide minerals interactions with sedimentary processes had no considerably decreases the iron oxide formation considerable influence on the PP concentration and P retention in the sediments (Canfield 1989, profile (Fig. 7a) of modeled sediment. However, Lehtoranta and Pitkänen 2003). Recall that the in the presence of a Beggiatoa mat, the modeled scheme in (Fig. 6) intends to specify the major sediment contained higher concentrations of Fe2+ processes and compounds controlling the mobil- and Fe oxide as compared with the Beggiatoa- ity of P in the sediments of the Baltic Sea. Thus, free sediment (Fig. 7b and c). These differ- we did not include less influential processes such ences between the two cases are attributable to as adsorption of P on other sedimentary minerals Beggiatoa sulfide consumption (Fig. 7d). In the or the formation of P-containing compounds. case of Beggiatoa-free sediment, sediment zones rich in iron oxides and iron ions (top sediment layers) are exposed to high concentrations of The case of Beggiatoa-covered sediment sulfide transferred upward from sulfidic layers and therefore the rates of Fe oxide dissolution In order to investigate the effects of Beggiatoa and Fe2+ mineralization by sulfide (Table 5: R7 metabolism on the P retention processes, we and R14, respectively) are high in upper sedi- hypothesized that the modeled sediment is cov- ment layers. 180 Yekta & Rahm • Boreal Env. Res. vol. 16

0 0 0 1 a 1 b 1 c 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7 Sediment depth (cm) 8 8 8 9 9 9 10 10 10 0 50 100 150 200 0 20 40 60 0 50 100 −1 PP (µM) Fe(II) (µM) Fe oxide (µmol g ) 0 0 0 1 d 0.5 e 0.5 f 2 1 1 3 1.5 1.5 4 2 2 Fig. 7. Comparison of simulated concentration 5 2.5 2.5 profiles of a( ) PP, (b) Fe2+, 6 3 3 (c) Fe oxide, (d) sulfide, 7 Sediment depth (cm) 3.5 3.5 (e) nitrate, and (f) oxygen 8 4 4 between the cases of Beg- 9 4.5 4.5 giatoa-free and Beggia- toa-covered sediments. 10 5 5 0 500 1000 0 1 2 3 0 1 2 3 Please note the different Sulfide (µM) Nitrate (µM) Oxygen (µM) scales of the profiles of Beggiatoa−free Beggiatoa−covered oxygen and nitrate.

The active iron oxide precipitation/dissolu- inside the sediment environment. tion cycle between oxic and anoxic layers of the Nitrate and oxygen consumption by the sediment promotes the local scale P adsorption/ Beggiatoa mat lowered the concentrations of desorption process (see Fig. 6). Contrary, iron these substances but did not inhibit their trans- oxide dissolution by sulfide and the following port from the water to the uppermost sediment iron sulfide formation withdraw iron molecules layers (Fig. 7e and f, respectively). Similarly, in from this cycle and decrease the supply of iron situ measurements of nitrate and oxygen con- active in terms of phosphate adsorption. Chemi- centration profiles by other researchers demon- cal dissolution of iron oxides by sulfide and strated the transport of these compounds to the precipitation of iron sulfide in top sediment top layers of the sediments occupied by Beggia- layers occur with lower rates in the presence of toa mats (Mußmann et al. 2003, Preisler et al. a Beggiatoa mat (Fig. 8). The model-calculated 2007). A Beggiatoa community in complex and Fe:TP ratio (summation of Fe oxide and dis- variable natural conditions might behave differ- solved Fe2+ on molar basis and the TP ratio) of ently from the presented model. For instance, 6.2:1 for the upper 2 cm of the Beggiatoa-cov- the model relies on the mechanism suggested ered sediment. This value is 63% higher in com- by Sayama et al. (2005) for defining Beggiatoa parison with the calculated Fe:TP ratio of 3.8:1 substrate uptake. These researchers observed a for Beggiatoa-free sediment under similar condi- complete shift from denitrification to “dissimi- tions. These results suggest that the separation latory ammonium production” in the presence of iron rich layers from the high concentrations of Beggiatoa filaments. However, other stud- of sulfide through Beggiatoa sulfide consump- ies suggested that Beggiatoa mats may perform tion increases the supply of iron available for denitrification rather than ammonium production the local scale P adsorption/desorption process (Sweerts et al. 1990). Hence, the role of Beggia- Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 181

0 0 a b 1 1

2 2

3 3

4 4

5 5

6 6

Sediment depth (cm) 7 7

8 8 Fig. 8. Comparison of (a) dissolution rate of Fe 9 9 oxide minerals by sulfide, and (b) Fe sulfide precipi- 10 10 tation rate in the presence 0 10 20 30 40 50 0 0.2 0.4 0.6 0.8 (solid lines) and absence Dissolution rate of Fe oxide by sulfide Fe sulfide precipitation rate (µmol g−1 yr−1) (µmol g−1 yr−1) (dashed lines) of Beggia- toa mat. Beggiatoa-free Beggiatoa-covered toa spp. in the N cycle in natural sediments results demonstrated potential local scale interac- might be different than what was predicted by tions of sulfide-oxidizing Beggiatoa with the P the current model. retention processes in their inhabited sediments. Furthermore, Beggiatoa mats on the sediments have a variety of densities which vary seasonally according to environmental condi- Implications for P management in the tions (e.g. Jørgensen 1977, Rosenberg and Diaz Baltic Sea 1993). Biomass density is an important factor for the impacts of Beggiatoa mats on the sedi- The amount of P released from newly formed ments (Preisler et al. 2007). Although we did not anoxic/hypoxic sediments of the Baltic Sea can specify the density of the Beggiatoa mat, bacte- be more than the riverine P loads which makes rial biomass was treated as a non-limiting factor P exchange across the sediment–water interface in uptake rates of the substrates representing the a critical process affecting both eutrophication abundance of Beggiatoa filaments in the mod- and hypoxia in this marine system (Conley et eled microbial community corresponding to a al. 2009). About a third of the Baltic Proper bot- microbial mat with a high biomass density. This toms were anoxic or hypoxic in the autumn 2008 assumption might lead to further overestimations (i.e. roughly all basins bellow 90 m, Hansson et of the impact of Beggiatoa when lacey cultures of al. 2008) and these conditions provide suitable these bacteria inhabit the sediment–water inter- habitat for the frequent occurrence of the genus face. The model did not specify sediment types Beggiatoa. The present model study highlighted and structure which are among other important the main processes regulating P retention in factors on physiological behavior of Beggia- these sediments and demonstrated that Beggia- toa spp. in marine sediments (Jørgensen 1977). toa spp., as one of the most probable surface spe- This simplification limits the applicability of the cies on hypoxic bottoms (Rosenberg and Diaz model when studying the Beggiatoa mats at dif- 1993), potentially interact with these processes. ferent Baltic Sea areas with different sediment There is currently a great deal of research physical characteristics. However, the numerical focused on remedy actions based on P retention experiment of this work was intended to study in the sediments to deal with eutrophication in the potential influences of Beggiatoa mats on the Baltic Sea (e.g the Swedish “BOX project” P retention processes. In this regard, the model and the Finnish “PROPPEN”). The P retention 182 Yekta & Rahm • Boreal Env. Res. vol. 16 capability of the sediment is limited when iron provided information on Beggiatoa physiology molecules are mineralized and buried as iron and enzymology, and additionally suggested that sulfides, but sulfide consumption by Beggiatoa these bacteria play an important role in the bio- mats is expected to have a positive contribution geochemistry of marine sediments. on P retention by increasing the supply of iron In this work, we studied the effects of Beggia- available for P adsorption/desorption in local toa metabolism on P retention using a diagenetic scale hypoxic sediments. Furthermore, internal model based on the dominant processes control- storage of nitrate by sulfide oxidizing Beggiatoa ling P mobility in the surficial sediments. In a is about four orders of magnitude greater than comparison of the model results using the case ambient concentrations which make the wide- of Beggiatoa-free sediment with available data, spread Beggiatoa filaments a potential sink for N we showed that the model is able to simulate in the Baltic Sea sediments (Rosenberg and Diaz realistic vertical concentration profiles of incor- 1993). Therefore, these bacteria may also play an porated compounds in the sediment–porewater important role in the regulation of the N stock in medium. We also showed that both the mobiliza- the Baltic Sea. tion and immobilization of P in the sediments To the best of our knowledge, there is no are mainly regulated by the extent of iron oxide comprehensive, large scale study focused on precipitation/dissolution and iron sulfide precipi- the spatiotemporal Beggiatoa distribution in the tation which induce the iron supply of the sedi- Baltic Sea. According to the presented results, ment to form reactive iron oxides or non-reactive we suggest that future remedy actions based on iron sulfides in terms of phosphate adsorption. P retention in the sediments of the Baltic Sea In one scenario for Beggiatoa-covered sedi- should consider the role of Beggiatoa mats and ment, we included the uptake rates of sulfide, its metabolism in sedimentary P retention proc- oxygen, and nitrate, as well as production rates esses. Information including detailed data on of ammonium and sulfate by the Beggiatoa mat the spatiotemporal distribution of these microor- in the chemical network of the modeled sedi- ganisms and their density as well as local scale ment. The results demonstrated that the interac- measurements of biogeochemical parameters tions of microbially mediated redox reactions controlling Beggiatoa interactions with sedimen- from the Beggiatoa metabolism with P retention tary processes is required in order to quantify the processes potentially increases the supply of iron contribution of these bacteria to the biogeochem- available for local scale P adsorption/desorp- istry and retention of P on both local and basin tion processes inside the sediment. This is done scales. This information and environmental data through the separation of iron rich layers from will be useful for the management of present P in deeper layers rich in sulfide as the Beggiatoa the Baltic Sea and future action planning. filaments consume sulfide in the top layers of the sediment. Finally, it seems that it would be beneficial Conclusions in the future to increase the knowledge of the biogeochemistry of Beggiatoa spp. in marine Ongoing eutrophication builds up P storage in systems in general and especially in the Baltic surficial sediments of the Baltic Sea comparable Sea where remedy actions based on P reten- to almost a decade of riverine P load. It is well tion in the sediments are under assessment. Our understood that shifts in bottom water conditions current study is a point of departure for future to hypoxia and anoxia result in both the mobi- investigations aiming to provide comprehensive lization of the IBP and the transfer of P to the information on the role of Beggiatoa spp. in the overlying water body. The portion of hypoxic/ biogeochemistry of the Baltic Sea, through the anoxic sediments has increased with time in the integration of the modeling approach with field Baltic Sea. These bottom conditions eliminate and laboratory measurements. higher orders of life but make the new sediment surfaces suitable habitats for sulfide-oxidizing Acknowledgements: Part of this study was done in conjunc- Beggiatoa. Previous studies by other researchers tion with the BOX-project and financed by the Swedish Boreal Env. Res. vol. 16 • Effects of Beggiatoa spp. on phosphorus in sediments 183

Environmental Protection Agency. We are most grateful for chim. Cosmochim. Acta 65: 2361–2378. the help offered by B.P. Boudreau regarding background Gao Y. & Mucci A. 2003. Individual and competitive adsorp- information. We would also like to acknowledge the con- tion of phosphate and arsenate on goethite in artificial structive comments from two anonymous referees of the seawater. Chem. Geol. 199: 91–109. article which helped to improve the manuscript. Geelhoed J.S., Hiemstra T. & Van Riemsdijk, W.H. 1997. Phosphate and sulfate adsorption on goethite: Single anion and competitive adsorption. Geochim. Cosmo- chim. Acta 61: 2389–2396. References Gunnars A., Blomqvist S., Johansson P. & Andersson C. 2002. Formation of Fe(III) oxyhydroxide colloids in Berner R.A. 1964. An idealized model of dissolved sulfate freshwater and brackish seawater, with incorporation of distribution in recent sediments. Geochim. Cosmochim. phosphate and calcium. Geochim. Cosmochim. Acta 66: Acta 28: 1497–1503. 745–758. Boudreau B.P. 1997. Diagenetic models and their implemen- Gustafsson B. & Stigebrandt A. 2007. Dynamics of tation: modeling transport and reactions in aquatic sedi- nutrients and oxygen/hydrogen sulfide in the Baltic ments. Springer-Verlag, Berlin. Sea deepwater. J. Geophys. Res. 112, G02023, Boudreau B.P. & Westrich J.T. 1984. The dependence of doi:10.1029/2006JG000304 bacterial sulfate reduction on sulfate concentration Hagen K.D. & Nelson D.C. 1997. Use of reduced sulfur in marine sediments. Geochim. Cosmochim. Acta 48: compounds by Beggiatoa spp.: Enzymology and physi- 2503–2516. ology of marine and freshwater strains in homogeneous Boudreau B.P. & Jørgensen B.B. 2001. The benthic boundary and gradient cultures. Appl. Environ. Microbiol. 63: layer transport processes and biogeochemistry. Oxford 3957–3964. University Press, Oxford. Hansson M., Thell A.-K., Andersson L. & Axe P. 2008. Brügmann L., Bernard P.C. & van Grieken R. 1992. Geo- Expeditionsrapport från U/F Argos: syrgaskartering i chemistry of suspended matter from the Baltic Sea 2. Östersjön. SMHI Dnr: Mo 2008-204. Results of bulk trace metal analysis by AAS. Mar. Chem. Hayduk W. & Laudie H. 1974. Vinyl chloride gas compress- 38: 303–323. ibility and solubility in water and aqueous potassium Canfield D.E. 1989. Reactive iron in marine sediments. Geo- laurate solutions. J. Chem. Eng. Data 19: 253–257. chim. Cosmochim. Acta 53: 619–632. Hille S., Nausch G. & Leipe T. 2005. Sedimentary deposi- Canfield D.E., Thamdrup B. & Hansen J.W. 1993. The anaer- tion and reflux of phosphorus (P) in the eastern Gotland obic degradation of organic matter in Danish coastal sed- Basin and their coupling with P concentrations in the iments: Iron reduction, manganese reduction, and sulfate water column. Oceanologia 47: 663–679. reduction. Geochim. Cosmochim. Acta 57: 3867–3883. Hille S., Leipe T. & Seifert T. 2006. Spatial variability of Carman R. & Rahm L. 1997. Early diagenesis and chemical recent sedimentation rates in the eastern Gotland Basin characteristics of interstitial water and sediments in the (Baltic Sea). Oceanologia 48: 297–317. deep deposition bottoms of the Baltic proper. J. Sea Res. Hinck S., Neu T.R., Lavik G., Mussmann M., De Beer D. 37: 25–47. & Jonkers H.M. 2007. Physiological adaptation of a Cloern J.E. 2001. Our evolving conceptual model of the nitrate-storing Beggiatoa sp. to diel cycling in a pho- coastal eutrophication problem. Mar. Ecol. Prog. Ser. totrophic hypersaline mat. Appl. Environ. Microbiol.73: 210: 223–253. 7013–7022. Conley D.J., Humborg C., Rahm L., Savchuk O.P. & Wulff F. Jensen H.S., Mortensen P.B., Andersen F.Ø., Rasmussen 2002. Hypoxia in the Baltic Sea and basin-scale changes E. & Jensen A. 1995. Phosphorus cycling in a coastal in phosphorus biogeochemistry. Environ. Sci. Technol. marine sediment, Aarhus Bay, Denmark. Limnol. Ocea- 36: 5315–5320. nogr. 40: 908–917. Conley D.J., Björck S., Bonsdorff E., Carstensen J., Destouni Jørgensen B.B. 1977. Distribution of colorless sulfur bacteria G., Gustafsson B.G., Hietanen S., Kortekaas M., Kuosa (Beggiatoa spp.) in a coastal marine sediment. Mar. Biol. H., Meier H.E.M., Müller-Karulis B., Nordberg K., 41: 19–28. Norkko A., Nürnberg G., Pitkänen H., Rabalais N.N., Jørgensen B.B. 1982. Ecology of the bacteria of the sul- Rosenberg R., Savchuk O.P., Slomp C.P., Voss M., Wulff phur cycle with special reference to anoxic-oxic inter- F. & Zillén L. 2009. Hypoxia-related processes in the face environments. Phil. Trans. R. Soc. Lond. B 298: Baltic Sea. Environ. Sci. Technol. 43: 3412–3420. 543–561. Diaz R.J. & Rosenberg R. 2008. Spreading dead zones Jørgensen B.B. & Revsbech N.P. 1983. Colorless sulfur

and consequences for marine ecosystems. Science. 321: bacteria, Beggiatoa spp. and Thiovulum spp., in O2

926–929. and H2S microgradients. Appl. Environ. Microbiol. 45: Fenchel T. & Bernard C. 1995. Mats of colourless sulphur 1261–1270. bacteria. I. Major microbial processes. Mar. Ecol. Prog. Kamp A., Stief P. & Schulz-Vogt H.N. 2006. Anaerobic Ser. 128: 161–170. sulfide oxidation with nitrate by a freshwater Beggia- Gao Y. & Mucci A. 2001. Acid base reaction, phosphate and toa enrichment culture. Appl. Environ. Microbiol. 72: arsenate complexation, and thier competitive adsorption 4755–4760. at the surface of geothite in 0.7 M NaCl solution. Geo- Kamp A., Røy H. & Schulz-Vogt H.N. 2008. Video-sup- 184 Yekta & Rahm • Boreal Env. Res. vol. 16

ported analysis of Beggiatoa filament growth, breakage, 525–544. and movement. Microb. Ecol. 56: 484–491. Rosenberg R. & Diaz R.J. 1993. Sulfur bacteria (Beggiatoa Krom M.D. & Berner R.A. 1980. The diffusion coefficients spp.) mats indicate hypoxic conditions in the Inner of sulfate, ammonium, and phosphate ions in anoxic Stockholm Archipelago. Ambio 22: 32–36. marine sediments, Long Island Sound. Limnol. Ocea- Ruttenberg K.C. & Berner R.A. 1993. Authigenic apatite nogr. 25: 327–337. formation and burial in sediments from non-upwelling, Lehtoranta J. & Pitkanen H. 2003. Binding of phosphate in continental margin environments. Geochim. Cosmo- sediment accumulation areas of the eastern Gulf of Fin- chim. Acta 57: 991–1007. land, Baltic Sea. Hydrobiologia 492: 55–67. Sayama M., Risgaard-Petersen N., Nielsen L.P., Fossing – Lehtoranta J., Ekholm P. & Pitkänen H. 2008. Eutrophica- H. & Christensen P.B. 2005. Impact of bacterial NO3 tion-driven sediment microbial processes can explain the transport on sediment biogeochemistry. Appl. Environ. regional variation in phosphorus concentrations between Microbiol. 71: 7575–7577. Baltic Sea sub-basins. J. Mar. Syst. 74: 495–504. Seifert T. & Kaiser B. 1995. A high resolution spherical grid Leipe T., Dippner J.W., Hille S., Voss M., Christiansen C. & topography of the Baltic Sea. Meereswissenschaftliche Bartholdy J. 2008. Environmental changes in the central Berichte 9 (Institut für Ostseeforschung, Warnemünde): Baltic Sea during the past 1000 years: inferences from 72–88. sedimentary records, hydrography and climate. Oceano- Shuler M.L. & Kargi F. 2002. Bioprocess engineering — logia 50: 23–41. basic concepts. Prentice Hall, New Delhi. McHatton S.C., Barry J.P., Jannasch H.W. & Nelson Sobolev D. & Roden E.E. 2001. Suboxic deposition of ferric D.C. 1996. High nitrate concentrations in vacuolate, iron by bacteria in opposing gradients of Fe(II) and autotrophic marine Beggiatoa spp. Appl. Environ. oxygen at circumneutral pH. Appl. Environ. Microbiol. Microbiol. 62: 954–958. 67: 1328–1334. Møller M.M., Nielsen L.P. & Jørgensen B.B. 1985. Oxygen Spiteri C., van Cappellen P. & Regnier P. 2008. Surface responses and mat formation by Beggiatoa spp. Appl. complexation effects on phosphate adsorption to ferric Environ.Microbiol. 50: 373–382. iron oxyhydroxides along pH and salinity gradients in Mortimer, C.H. 1941. The exchange of substances between estuaries and coastal aquifers. Geochim. Cosmochim. mud and water in lakes. I. J. Ecol. 29: 280–329. Acta 72: 3431–3445. Mußmann M., Schulz H.N., Strotmann B., Kjær T., Nielsen Sternbeck J. & Sohlenius G. 1997. Authigenic sulfide and L.P., Rosselló-Mora R.A., Amann R.I. & Jørgensen carbonate mineral formation in Holocene sediments of B.B. 2003. Phylogeny and distribution of nitrate-storing the Baltic Sea. Chem. Geol. 135: 55–73. Beggiatoa spp. in coastal marine sediments. Environ. Sweerts J.-P.R.A., de Beer D., Nielsen L.P., Verdouw H., van Microbiol. 5: 523–533. den Heuvel J.C., Cohen Y. & Cappenberg T.E. 1990. Myers C.R. & Nealson K.H. 1988. Microbial reduction of Denitrification by sulphur oxidizingBeggiatoa spp. mats manganese oxides: Interactions with iron and sulfur. on freshwater sediments. Nature 344: 762–763. Geochim. Cosmochim. Acta 52: 2727–2732. Tejedor-Tejedor M.I. & Anderson M.A. 1990. Protonation Nelson D.C., Jørgensen B.B. & Revsbech N.P. 1986. Growth of phosphate on the surface of goethite as studied by pattern and yield of a chemoautotrophic Beggiatoa sp. CIR-FTIR and electrophoretic mobility. Langmuir 6: in oxygen-sulfide microgradients. Appl. Environ. Micro- 602–611. biol. 52: 225–233. Tromp T.K., van Cappellen P. & Key R.M. 1995. A global Persson P., Nilsson N. & Sjöberg S. 1996. Structure and model for the early diagenesis of organic carbon and bonding of orthophosphate ions at the iron oxide-aque- organic phosphorus in marine sediments. Geochim. Cos- ous interface, J. Colloid Interface Sci. 177: 263–275. mochim. Acta 59: 1259–1284. Preisler A., de Beer D., Lichtschlag A., Lavik G., Boetius A. Tsai D.S. & Strieder W. 1986. Radiation across and down a & Jørgensen B.B. 2007. Biological and chemical sulfide cylindrical pore having both specular and diffuse reflect- oxidation in a Beggiatoa inhabited marine sediment. ance components. Ind. Eng. Chem. Fundam. 2: 244–249. ISME J. 1: 341–353. Vallius H. 2006. Permanent seafloor anoxia in coastal basins Redfield A. 1958. The biological control of chemical factors of the northwestern Gulf of Finland, Baltic Sea. Ambio in the environment. Am. Sci . 46: 205–221. 35: 105–108. Rickard D. 1995. Kinetics of FeS precipitation: Part 1. Van Cappellen P. & Wang Y. 1996. Cycling of iron and Competing reaction mechanisms. Geochim. Cosmochim. manganese in surface sediments: A general theory for Acta 59: 4367–4379. the coupled transport and reaction of carbon, oxygen, Robertson L.A. & Kuenen J.G. 1984. Aerobic denitrification nitrogen, sulfur, iron, and manganese. Am. J. Sci. 296: — old wine in new bottles? Anton. Leeuw. Int. J. G. 50: 197–243.