SEARCH

SEARCH BY CITATION

Keywords:

  • biogeochemical turnover;
  • hot spots;
  • micro-topography;
  • modeling;
  • wetland

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] Wetlands provide important ecohydrological services by regulating fluxes of nutrients and pollutants to receiving waters, which can in turn mitigate adverse effects on water quality. Turnover of redox-sensitive solutes in wetlands has been shown to take place in distinct spatial and temporal patterns, commonly referred to as hot spots and hot moments. Despite the importance of such patterns for solute fluxes the mechanistic understanding of their formation is still weak and their existence is often explained by variations in soil properties and diffusive transport only. Here we show that surface micro-topography in wetlands can cause the formation of biogeochemical hot spots solely by the advective redistribution of infiltrating water as a result of complex subsurface flow patterns. Surface and subsurface flows are simulated for an idealized section of a riparian wetland using a fully integrated numerical code for coupled surface-subsurface systems. Biogeochemical processes and transport along advective subsurface flow paths are simulated kinetically using the biogeochemical code PHREEQC. Distinct patterns of biogeochemical activity (expressed as reaction rates) develop in response to micro-topography induced subsurface flow patterns. Simulated vertical pore water profiles for various redox-sensitive species resemble profiles observed in the field. This mechanistic explanation of hot spot formation complements the more static explanations that relate hot spots solely to spatial variability in soil characteristics and can account for spatial as well as temporal variability of biogeochemical activity, which is needed to assess future changes in the biogeochemical turnover of wetland systems.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] Wetlands provide important ecohydrological services in many mountainous headwater catchments. They store significant amounts of carbon as peat, and act as effective nutrient sinks e.g., for sulfur, phosphorus and nitrogen [Kellogg and Bridgham, 2003; Paul et al., 2006; Tauchnitz et al., 2010]. Redox conditions and the corresponding biogeochemical processes in these wetlands largely control the source and sink functions of peat-soil dominated catchments [Bishop et al., 2004; Lischeid et al., 2007]. Process activities in such wetlands are spatially nonuniform, though, and have been found to form distinct hot spots [Jacks and Norrström, 2004], i.e. areas or patches that show disproportionally high reaction rates relative to the surrounding areas [McClain et al., 2003; Morris and Waddington, 2011]. Such hot spots are not easily identified in the scatter of spatiotemporal data sets and hence their relevance for net matter turnover is assumed to be underestimated [Richardson et al., 2007; McClain et al., 2003; Vidon et al., 2010]. Various studies have observed large variations in the spatial distribution of redox sensitive solutes within wetland soils [Jacks and Norrström, 2004; McMahon and Chapelle, 2008] on the scale of transects (10–50 m) [Jacks and Norrström, 2004] as well as in the meter and sub-meter range [Knorr and Blodau, 2009; Mitchell and Branfireun, 2005; Wachinger et al., 2000]. It seems obvious that complex transport and transformation processes within the subsurface are main drivers for the observed spatial heterogeneity in solute concentrations. Although studies have pointed at potential effects of subsurface flow dynamics in wetlands on solute concentrations, e.g., by enhanced mixing due to hydraulic gradient reversals [Reeve et al., 2006] and the formation of hot spots has conceptually been linked to transport processes [McClain et al., 2003] transport and biogeochemical transformations are rarely combined mechanistically to explain such phenomena. Recent studies in wetlands have mainly attributed the formation of hot spots to lateral variations in local physico-chemical variables such as soil texture, composition, moisture or temperature [Bruland and Richardson, 2005; Morris and Waddington, 2011] or the local availability of certain reactants such as nitrate or DOC [Bruland et al., 2006]. Differences in these properties may e.g., arise from different degrees of peat decomposition, peat compaction, vegetation or surface micro-topography [Gafni and Kenneth, 1990; Cheng et al., 2011; Bruland and Richardson, 2005].

[3] This perspective, however, does not consider that microbial processes are dynamic and dependent on variable hydrologic and biogeochemical boundary conditions. The close links between the mechanisms controlling biogeochemical activity in wetlands and the hydrological processes occurring within the wetland have been highlighted in several studies [Morris and Waddington, 2011; Mitchell and Branfireun, 2005]. Field studies [Knorr et al., 2009; Knorr and Blodau, 2009] demonstrated a rapid change of predominant redox processes (i.e., iron(III)-, sulfate reduction and methanogenesis) in a wetland exposed to fluctuations of hydrological boundary conditions during manipulation of the water level. Wetlands in mountainous catchments are often characterized by rapidly fluctuating but shallow water levels [Devito and Hill, 1997; Lischeid et al., 2007]. Such hydrological conditions facilitate fast flow components like saturation excess overland flow and shallow subsurface flows [Frei et al., 2010; Holden and Burt, 2003]. The dynamics of these flow components are important controls on mobilization of dissolved solutes (like e.g., dissolved organic carbon or nitrate) from wetlands [Alewell et al., 2007; Lischeid et al., 2007; Hinton et al., 1998; Dosskey and Bertsch, 1994] but their effect on the biogeochemical processes and distribution of redox sensitive solutes is still poorly understood and rarely addressed [Shabaga and Hill, 2010]. Partly this is because it is nearly impossible to directly investigate and characterize the complex, dynamic subsurface hydrology in the field. Therefore the interpretation of field observations (e.g., depth profiles for redox sensitive solutes) may be poorly constrained, e.g., if biogeochemical turnover rates are calculated based on the assumption that resupply of dissolved electron acceptors/donors within riparian wetlands is only diffusion limited [Beer and Blodau, 2007; Clymo and Bryant, 2008]. This simplification may hold true for some sites [Beer and Blodau, 2007] and for defined lab incubations [Knorr and Blodau, 2009], but it neglects that transport and turnover of redox sensitive solutes at many natural sites occurs within a complex, three dimensional (3D) subsurface flow field that is subject to variable boundary conditions. This results in distinct flow paths along which biogeochemical reactions can occur, controlled by the individual kinetics of each process [Knorr and Blodau, 2009; Hill, 2000; Brovelli et al., 2011]. An improved mechanistic model for the formation and occurrence of biogeochemical hot spots therefore needs to account for flow and transport processes and how they are affected by changes in hydrologic boundary conditions. This is of particular importance if such a model is used to assess the effects of climate change where induced shifts in the frequency of intense rainstorms or extended droughts [Huntington, 2006] have the potential to significantly alter the boundary conditions within wetlands.

[4] Virtual experiments [Weiler and McDonnell, 2004, 2006] have proven to be a suitable tool to investigate complex hydrologic processes and feedback mechanisms between hydrology and biogeochemistry [Frei et al., 2010; Boano et al., 2010; Jakobsen, 2007]. In this study, we use virtual modeling experiments to investigate how complex subsurface flow patterns induced by surface micro-topography affect the subsurface transport of redox sensitive solutes and the resulting spatial distribution of biogeochemical process activities within a hummocky wetland. We test the hypothesis that the complex subsurface flow field creates biogeochemical conditions in the subsurface that facilitate the formation of local process hot spots even in soils with uniform soil properties. To address this objective, the numerical simulations of complex surface and subsurface flow processes in the hypothetical section of the riparian wetland with pronounced micro-topography (hollows and hummocks) as described byFrei et al. [2010], is combined with advective particle tracking and multispecies biogeochemical simulations in a sequential stream tube approach. The main redox reactions typically found in peat-forming wetlands are simulated along individual subsurface flow paths, which are subject to local changes in the biogeochemical boundary conditions, using the geochemical model PHREEQC [Parkhurst, 1995]. The simulated wetland reflects the structural and hydrological characteristics of a riparian wetland in the Lehstenbach catchment in South-East Germany [Paul et al., 2006] which are not uncommon for peat-forming wetlands elsewhere [e.g.,Holden and Burt, 2003; Inamdar et al., 2009]. Hence this study will improve our general understanding of how subsurface hydrology affects redox transformations and turnover rates within riparian wetlands.

2. Material and Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

2.1. Surface/Subsurface Flow Simulation and Particle Tracking

[5] The numerical flow model presented by Frei et al. [2010]for a small hypothetical section of a wetland with micro-topography (10 m by 20 m with a maximum thickness of 2 m) draining into a channel segment (seeFigure 3) was used to simulate riparian runoff generation and subsurface flow patterns. The flow model and its parameterization are only briefly summarized here, as the model is described in more detail in Frei et al. [2010]. The hummocky micro-topography is geostatistically simulated based on a Markov Chain model of transition probabilities, which is derived and conditioned with surveyed elevation data from the Lehstenbach field site [Frei et al., 2010; Carle and Fogg, 1996]. Two different models of micro-topography, using a mean length of 0.5 m and 0.25 m respectively, were generated to represent differently sized hollow and hummock structures. Micro-topography realizations were then superimposed on top of a planar, slightly inclined surface (slope = 0.03) to create a realistic representation of a typical wetland section at the field site as a basis for the flow model (for details seeFrei et al. [2010]). Micro-topography models were compared to a model with a planar surface as a reference (hereafter referred to as ‘planar reference model’). Transient surface and subsurface flow were simulated using the numerical code Hydrogeosphere (HGS) [Therrien et al., 2008] which provides a fully integrated 3D solution for variably saturated subsurface flow (Richards equation) and a 2D depth-averaged solution for surface flows based on the diffusive wave approximation to the St. Venant equations [Therrien et al., 2008]. HGS is increasingly used for the simulations of coupled surface-subsurface hydrologic systems [e.g.,Brookfield et al., 2009; Jones et al., 2006]. To drive the flow models observed daily precipitation for the hydrologic year (HY) 2000 (10/31/1999–11/1/2000), which represents typical hydrometeorological conditions in the Lehstenbach catchment, were applied as a flux boundary at the model surface. To simplify data handling (each model output file contains velocity data for about 210,000 model nodes) transient model output was only generated in five day intervals (integration time step Δt = 5d).

[6] Subsurface flow paths for each of the flow models (planar reference + 2 realizations of micro-topography) were derived by applying an advective particle tracking routine, implemented in the Tecplot 360 post-processing software [Tecplot, Inc., 2003], to a transient flow field that consists of a multiyear (25 years) sequence of the velocity fields simulated for the reference year (HY 2000). 21,000 particles (one per surface node) were placed on the model surface yielding 21,000 individual subsurface flow paths. Particles were tracked from infiltration until the particle leaves the subsurface domain due to exfiltration.

2.2. Multispecies Biogeochemical Simulations

2.2.1. Coupling Hydrology and Biogeochemistry

[7] The basic concept of implementing a biogeochemical model along individual subsurface flow paths is illustrated in Figure 1. Every flow path is split into imax[−] different sub-sections where imax represents how often the integration time step Δt = 5d is being repeated until the water particle leaves the subsurface domain due to exfiltration. For a known flow path (as shown in Figure 1) the total subsurface residence time (RTtotal [T]) for a particle traveling along that path (from the moment of its infiltration until exfiltration) can be approximated by equation (1):

  • display math
image

Figure 1. Schematic plot showing how subsurface hydrology was coupled to the biogeochemical model PHRREQC by using particle tracking techniques. Isolated sub surface flow pathlines are split into individual sub-sections for which the redox chemical conditions are simulated depending on the hydrological and biogeochemical boundary conditions.

Download figure to PowerPoint

[8] According to Figure 1, each of the different sub-sectionsi has a start (xi-1, yi-1, zi-1) [L, L, L] and end (xi, yi, zi) location [L, L, L], a sub-section's residence time (RTsub[T]) representing the time a water particle spends within a sub-sectioni[-] and a characteristic travel distance (di[L]). Each sub-section has a constant sub-section residence time (RTsub) of five days. The flow distance di a water particle travels during RTsubdepends on the subsurface flow field which varies along the flow path according to the transient solution of the numerical flow model. The characteristic travel distance for a sub-section (di) can be linearly approximated via the sub-section's start (xi-1, yi-1, zi-1) and end locations (xi, yi, zi):

  • display math

The total distance D[L] a particle travels from the moment of infiltration until exfiltration can be approximated by summing up the individual travel distances for all sub-sections di:

  • display math

For a single subsurface flow path, the corresponding biogeochemical simulation consists of imax individual PHREEQC [Parkhurst, 1995] scripts, one script for each sub-section of a subsurface flow path (lower table ofFigure 1). Each script uses different boundary conditions which were individually derived from the numerical flow model's solution. During the integration time step (Δt = 5d), the boundary conditions for a single PHREEQC [Parkhurst, 1995] sub-section are kept constant (detailed information on the boundary conditions are given in the following section). A single script simulates all relevant redox reactions, implemented as kinetic formulations (detailed information on the implemented reactions is also given in a subsequent paragraph), that occur within the sub-section during the integration time step Δt. For all of the 21,000 flow paths and for each of the flow models this approach was used resulting in about 1,450,000 different PHREEQC [Parkhurst, 1995] sub-section simulations per flow model.

2.2.2. Initial and Boundary Conditions

[9] With the exception of the first sub-section of each flow path (representing the first 5 days after infiltration), all PHREEQC [Parkhurst, 1995] sub-section simulationsiuse the final chemical conditions of the previous sub-section (FCi-1) simulation as an initialization for the subsequent sub-section simulation (Figure 1, lower table). The first sub-section simulation of each flow pathi = 0 uses uniformly assigned initial conditions (ICglobal), which are listed in Table 1. Values for ICglobal are based on the chemical composition and pH of shallow pore water at the wetland field site in the Lehstenbach catchment [Knorr et al., 2009]. Some redox-reactions within the PHREEQC [Parkhurst, 1995] simulations only occur under the presence of oxygen (like e.g., aerobic respiration or different oxidation processes). Other processes like e.g., denitrification, iron and sulfate reduction are only initiated under conditions where oxygen concentrations are low or zero. Along a flow path, availability of oxygen changes depending on the hydrological conditions (e.g., flow in saturated versus unsaturated media): Within the unsaturated zone, availability of oxygen is assumed to be unlimited because depleted oxygen is continuously supplied by atmospheric diffusion with the rate of resupply (τresupply) exceeding the rate of oxygen depletion (τdepletion). In the saturated zone, where pores are completely saturated, water acts as an effective diffusion barrier and oxygen becomes increasingly limiting with growing depth below the water table. At a certain depth in the saturated zone τresupply becomes equal to τdepletion. Below this point depletion exceeds supply and no more oxygen is available. Availability of oxygen was thus used as the key control for either initiating or suppressing the series of anaerobic redox processes in the biogeochemical simulations. Availability of oxygen, as a boundary condition for the PHREEQC [Parkhurst, 1995] sub-section simulations, was coupled to the transient pressure heads phi-1 obtained from the flow model, as these pressure heads describe the relative position with respect to the current local water table (i.e., saturated or unsaturated zone). Coupling oxygen availability to transient pressure heads was performed as follows: (1)For each sub-section's start location (xi-1, yi-1, zi-1) the corresponding time dependent pressure head (phi-1) was estimated from the solution of the numerical flow model. (2) The pressure head (phi [L]) was used as an indicator for three different conditions or zones according to Figure 2. Zone 1 (negative pressure head) represents the unsaturated zone where significant fractions of the soil matrix pores are air filled and where oxygen content is constantly high and in equilibrium with the atmosphere. Zone 2 (0 m < phi< 0.25 m) represents the transition zone between the zone saturated with oxygen (zone 1) and deeper water-saturated layers where oxygen is completely depleted (zone 3). Within zone 2, atmospheric diffusion becomes less effective, in terms of resupply, with increasing pressure heads (indicative of increasing depths below the water table). If the pressure head phi is located within either zone 1 or 2, the oxygen boundary condition (BCi) is assigned according to the oxygen - pressure head relationship shown inFigure 2, which was derived from observed depth profiles of oxygen sensitive redox species taken at the field site [Knorr et al., 2009]. During the sub-section integration time step Δt, it is assumed that the assigned boundary conditions do not change, which means that the corresponding oxygen concentration during a sub-section simulation remains constant for sub-sections located within zone 1 and 2. For sub-sections where the corresponding pressure head phi is located within zone 3 of Figure 2, oxygen is not assigned as a boundary condition instead the PHREEQC [Parkhurst, 1995] simulation for these sub-sections is initialized with an oxygen content equal to the residual oxygen content of the preceding sub-section simulationi-1. For sub-sections where the corresponding pressure head phi lies within zone 3, oxygen can be depleted by oxygen consuming redox reactions. All represented reductive processes (detailed information are given in the next paragraph) are treated as reactions catalyzed by microorganisms, comparable to e.g., the process model for methane production in wetlands as shown by Segers and Kengen [1998]. These types of reactions depend on the presence of (a) an adequate electron acceptor (like e.g., oxygen, nitrate, iron(III) or sulfate) and (b)a source of labile carbon that is available to microorganisms. As a simplification to reduce model complexity, we considered the electron acceptor as the limiting factor for the presence of the individual catalyzed redox-reactions. We think that this is a reasonable approximation, since the supply of labile carbon (e.g., acetate) may be assumed to be coupled to the organic matter mineralization rate, as usually no intermediates (e.g., from fermentation) accumulate [Segers and Kengen, 1998]. The biogeochemical simulations were thus performed based on that concept, implementing an unlimited carbon source as BC for all sub-section simulations and limiting process rates solely by their kinetic parameters. By dynamically assigning the biogeochemical boundary conditions to each individual sub-section the whole sequence of sub-section simulations for one subsurface flow path, can be viewed as a continuous simulation of the redox-chemical evolution of a small water parcel that carries dissolved redox sensitive solutes and is transported along that specific flow path.

Table 1. Initial Concentrations ICglobal for Redox Sensitive Speciesa
SpeciesICglobal
  • a

    ICglobalis used to initialize the first PHREEQC sub-section simulation of each flow path. Values are based on field data determined for pore water, near to the surface of a typical wetland in the Lehstenbach catchment [Knorr et al., 2009].

Ammonium0 mol/L
Nitrate1.63 × 10−6 mol/L
Fe(III)6.76 × 10−5 mol/L
Fe(II)1.80 × 10−6 mol/L
Sulfate3.95 × 10−5 mol/L
Sulfide0 mol/L
Ph4.5
image

Figure 2. Typical oxygen depth profile based on observations from a riparian wetland site in the Lehstenbach catchment. Profile was used to assign oxygen boundary conditions to the different PHREEQC sub-section simulations based on transient flow model output.

Download figure to PowerPoint

2.2.3. Implemented Reactions and Kinetics

[10] For each sub-section, PHREEQC simulates redox processes as kinetic reactions based on the assigned boundary (BCi) and initial conditions (ICglobal/FCi-1). Implemented processes are shown in Table 2. All reduction processes are formulated based on Monod kinetic reactions according to equation (4):

  • display math

Here Rk [ML−3 T−1] is the kinetic rate of the corresponding reduction reaction k (k ∈ [1, 2, 3, 4]) according to Table 2. μmax,k [ML−3 T−1] represents the maximal specific growth rate (for k = 1 aerobic respiration, k = 2 de-nitrification, k = 3 iron(III)-reduction, k = 4: sulfate reduction) and Ks,k [ML−3] represents the substrate saturation constant (i.e., substrate concentration for k = 1 oxygen, k = 2 nitrate, k = 3 iron(III), k = 4 sulfate at half μmax, k). Ck [ML−3] is the corresponding concentration of the electron acceptor (for k = 1: oxygen, k = 2: nitrate, k = 3: iron(III); k = 4: sulfate). Monod kinetic coefficients (μmax, k and Ks,k) for all reduction processes are based on values reported for biodegradation of organic chemicals in aquifers [Appelo and Postma, 2005; Bekins et al., 1998; Schirmer et al., 1999; MacQuarrie et al., 1990; Eckert and Appelo, 2002; Kelly et al., 1996; Goldsmith and Balderson, 1988] and were later modified and adjusted as part of the calibration process. Simulated depth profiles for redox sensitive compounds (nitrate, sulfate and iron(II)) were calibrated by systematic variation of the Monod coefficients to best fit observed data taken at the study site [Knorr and Blodau, 2009; Knorr et al., 2009]. Calibrated Monod coefficients are listed in Table 2. For all processes where organic carbon is being decomposed, organically bound nitrogen is being released according to the Redfield ratio [Redfield, 1934]. Oxidation processes (k = 5 iron(II) oxidation, k = 6 nitrification, k = 7 aerobic sulfide oxidation and k = 8 anaerobic sulfide oxidation) were formulated using higher order reaction kinetics as listed in Table 2.

Table 2. Implemented Processes and the Equivalent Reaction Specific Kinetic Ratea
ProcessRateCoefficientsReference
  • a

    Reduction processes are formulated based on Monod type reaction kinetics.

Aerobic respirationaccording to equation (4)μmax,1 = 1.6 × 10−9 mol/Ls, Ks,1 = 2.9 × 10−6 mol/Lmodified and calibrated after Appelo and Postma [2005], Bekins et al. [1998], Schirmer et al. [1999], MacQuarrie et al. [1990], Eckert and Appelo [2002], Kelly et al. [1996], and Goldsmith and Balderson [1988]
Denitrificationaccording to equation (4)μmax,2 = 1.06 × 10−9 mol/Ls, Ks,2 = 2.0 × 10−6 mol/Lmodified and calibrated after Appelo and Postma [2005], Bekins et al. [1998], Schirmer et al. [1999], MacQuarrie et al. [1990], Eckert and Appelo [2002], Kelly et al. [1996], and Goldsmith and Balderson [1988]
Iron(III) reductionaccording to equation (4)μmax,3 = 1.5 × 10−12 mol/Ls, Ks,3 = 2.94 × 10−6 mol/Lmodified and calibrated after Appelo and Postma [2005], Bekins et al. [1998], Schirmer et al. [1999], MacQuarrie et al. [1990], Eckert and Appelo [2002], Kelly et al. [1996], and Goldsmith and Balderson [1988]
Sulfate reductionaccording to equation (4)μmax,3 = 0.5 × 10−10 mol/Ls, Ks,4 = 2.5 × 10−6 mol/Lmodified and calibrated after Appelo and Postma [2005], Bekins et al. [1998], Schirmer et al. [1999], MacQuarrie et al. [1990], Eckert and Appelo [2002], Kelly et al. [1996], and Goldsmith and Balderson [1988]
Iron(II) oxidation inline imageA5 = 8 × 1013 min−1 atm−1Appelo and Postma [2005] and Stumm and Morgan [1995]
Ammonium oxidation inline imageA6 = 5 × 106 (mol/L)−1 a−1Billen [1982] and van Cappellen and Wang [1996]
Aerobic sulfide oxidation inline imageA7 = 1.6 × 105 (mol/L)−1 a−1Millero et al. [1987] and van Cappellen and Wang [1996]
Anaerobic sulfide oxidation inline imageA8 = 8 × 103 (mol/L)−1 a−1Pyzik and Sommer [1981] and van Cappellen and Wang [1996]

[11] In redox controlled systems like wetlands, reduction processes can be expected to occur sequentially due to thermodynamic reasons [e.g., Achtnich et al., 1995]. Oxygen is used as primary electron acceptor, and after depletion nitrate, subsequently iron(III) and finally sulfate are being reduced. Further electron acceptors, such as manganese [Nealson and Saffarini, 1994] or organic molecules [Lovley et al., 1996] were not considered in this study. To make sure that the reduction processes proceed sequentially in the biogeochemical simulations, specific redox conditions were defined. These conditions are represented by critical concentrations for redox sensitive solutes which control whether a redox process can be initiated or not. Critical concentrations Ccrit [ML−3] for oxygen, nitrate and Iron(III) were derived based on observed depth profiles for redox sensitive compounds [Knorr and Blodau, 2009; Knorr et al., 2009; Estop-Aragonés and Blodau, 2012] For example, the critical concentration for oxygen CcritO2 is the residual concentration of oxygen under which denitrification is being initiated, which was estimated from observed depth profiles and field data. Critical concentrations for oxygen, nitrate and iron(III) are listed in Table 3. The rows of Table 3represent the conditions under which the different reduction processes are initiated. Entries must be read row-wise, where entries “>0” mean that the corresponding redox sensitive reactant (column) must be present and “-” means that this process is independent from the presence of this specific compound. For example iron(III) reduction in the biogeochemical simulation is initiated if:(1) Dissolved oxygen concentrations fall below CcritO2; (2) Most of the nitrate is already depleted where concentrations for nitrate fall below CcritNO3; (3) The electron acceptor iron(III) is available. Intervals for the activation of reduction processes are overlapping which means that multiple processes can occur simultaneously in the simulation; this was also observed in laboratory and under field conditions [Knorr and Blodau, 2009; Knorr et al., 2009].

Table 3. Critical Concentrations Which Are Controlling the Sequential Initialization of the Redox Sequencea
 OxygenNitrateIron(III)Sulfate
  • a

    Values were derived from field observations. Table must be read row wise (e.g., denitrification is initiated if 1. Oxygen contents drop below Ccrit derived for oxygen and 2. if nitrate is present). CcritO2 = 5.0 × 10−6 mol/L, CcritNO3 = 4.0 × 10−7 mol/L, and CcritFe3+ = 5.0 × 10−6 mol/L.

Aerobic respiration>0---
Denitrification<CcritO2>0--
Iron(III) reduction<CcritO2<CcritNO3>0-
Sulfate reduction<CcritO2<CcritNO3<CcritFe3+>0
2.2.4. Simplifying Model Assumptions

[12] To reduce the complexity of the represented system and to maintain a tractable model the following simplifying assumptions were made: (1)Soil specific parameters (saturated hydraulic conductivity, porosity and retention curves for variably saturated flow) are uniform within the model domain to separate the effects of micro-topography on subsurface flow dynamics from possible impacts of heterogeneity.(2) By simulating biogeochemical reactions along isolated subsurface flow paths, it is assumed that there is no interaction between different flow paths where water and/or solutes are exchanged due to hydrodynamic dispersion (mechanical dispersion + molecular diffusion). (3)Subsurface flow paths are derived based on a transient flow field resulting from yearly model runs. Particle tracking is performed for a 25 year period by repeating the yearly output of the flow model twenty-five times. This assumes that there are no inter-annual changes in the basic properties of the subsurface flow field (distribution of flow paths and RTs).(4)In the biogeochemical simulations availability of DOC, as the primary electron source for microbially catalyzed reactions (aerobic respiration, denitrification, iron(III)- and sulfate reduction) was assumed to be non-limiting.(5) Effects of vegetation and its potential influence on subsurface flow and redox processes, i.e., due to root respiration or exudation and evapotranspiration, are not considered. (6) Iron(III) species in the biogeochemical simulations are treated as solutes only, which are advectively transported within the subsurface domain and not as immobile solids bound to the peat matrix. (7) In the biogeochemical simulations, bioavailability of all involved species is not affected by e.g., complexation with DOC.

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

3.1. Subsurface Flow Patterns

[13] Subsurface flow paths for the two micro-topography models and the planar reference model are shown inFigures 3a–3d. For the model with a mean-length of 0.5 m (ml-0.5 m) flow paths are shown for the entire 3D model domain (Figure 3a) as well as for the 2D transect located across the center of the 3D model domain (dashed line in Figure 3a). The 2D flow fields for the transects represent projections of the 3D flow paths into a 2D plane (flow components in the y directions are neglected). In contrast to the planar reference model, both micro-topography models showed complex distributed subsurface flow paths where coexisting shallow and deep flow cells developed in 3D. This is a common phenomenon caused by topography and was first described byTóth [1962] for regional groundwater flow systems but can be found for flows in systems with pronounced topography over a range of scales [Wörman et al., 2006; Stonedahl et al., 2010]. Shallow flow cells are most pronounced for the model with a mean length of the surface structures of 0.5 m (Figure 3b) and are associated with the dominant surface structures (largest hummocks). Areas characterized by shallow flow cells are outlined with red dotted lines in Figure 3b. Water infiltrating in these areas relatively quickly returns to the land surface, travels shorter distances and is characterized by short subsurface residence times (Figures 4a and 4b). In contrast deeper flow cells, which develop for areas where water infiltrates deep into the subsurface predominately at locations that are located far away from the channel segment, have longer travel distances (often spanning the entire extent of the model domain) and show significantly longer residence times as also reflected in the water ages (residence time in the subsurface since infiltration) plotted for the central 2D transect in Figure 5b. Deeper flow cells are controlled by the general hydraulic gradient across the model domain. The flow field for the micro-topography model with a mean length of 0.25 m (ml-0.25 m) shows no clear separation between shallow and deep flow cells because the topographic variations are too small to create sufficient variations in subsurface hydraulic potentials that could induce significant shallow flow cells (Figure 3c). Similarly in the planar reference model (Figure 3d) flow paths are relatively uniform in space with flow directions almost parallel to the planar land surface.

image

Figure 3. Subsurface flow paths derived from particle tracking. (a) Flow paths for the 3D domain of the micro-topography realization with a mean length of 0.5 m. (b–d) Flow paths projected to a cross section at the center of the 3D domain (yellow dashed line in Figure 3a) for the two micro-topography models and the planar reference model. Outlined areas (red dotted lines) in Figure 3b represent the typical down and upwelling movement of the shallow flow system induced by surface micro-topography. Yellow dotted lines represent two flow paths (infiltrating at X = 0.4 m and X = 6.8 m) reflecting long and short subsurface residence times for which the biogeochemical evolution is shown inFigure 4. The model domain is 10 m × 20 m × 2 m.

Download figure to PowerPoint

image

Figure 4. Results of biogeochemical simulations along two different flow paths. (a, c, e, g) A deep flow path with long subsurface residence time and (b, d, f, h) a flow path of the shallow flow system. Subsurface flow velocities, pressure heads, flow depths and travel distances as shown in Figures 4a–4d were derived from numerical flow modeling and were used as hydrologic boundary conditions for the biogeochemical simulations. Additionally, oxygen availability (Figures 4e and 4f) was coupled to the pressure head dynamics (Figures 4c and 4d), individually for each flow path. Figures 4e and 4f show the evolution for oxidized species (nitrate, iron(III) and sulfate) in time normalized to their corresponding initial concentrations Ct=0 and Figures 4 g and 4 h the evolution of reduced species (ammonium, iron(II) and sulfide) normalized to their final concentrations Ct = max. How redox conditions are changing in time is also shown in Animations 1 and 2.

Download figure to PowerPoint

image

Figure 5. Results of the biogeochemical simulations shown for the sulfate reduction/oxidation process of the micro-topography scenario with the mean length 0.5 m. (a) Flow paths along which PHREEQC simulations were performed. Results were interpolated into the 2D cross sections. (b) The age distribution in years of subsurface flow derived from backward particle tracking. (c and d) Process activity of sulfate reduction/oxidation (kinetic rate in mol/Ls). Concentrations in mol/L for (e) sulfate and (f) sulfide.

Download figure to PowerPoint

3.2. Biogeochemical Evolution Along Flow Paths

[14] Figures 4e–4hdepict exemplarily the results of the biogeochemical simulations, shown for two selected subsurface flow paths of the ml-0.5 m micro-topography model. The two flow paths, beginning at location X = 0.4 m and X = 6.8 m (shown as yellow dotted lines inFigure 3b), represent the deep and shallow flow cells respectively. Results for the deep flow path are shown in Figures 4a, 4c, 4e, and 4g and for the shallow one in Figures 4b, 4d, 4f, and 4h. Both flow paths start in the unsaturated zone where pressure heads are negative (Figures 4c and 4d). For the unsaturated zone, dissolved oxygen concentrations are constantly high (Figures 4e and 4f) due to unlimited diffusive supply of atmospheric oxygen. Aerobic respiration is the dominant process within the unsaturated zone. The high turnover of organic material and the associated release of organically bound nitrogen within the unsaturated zone results in increasing concentrations of ammonium (Figures 4g and 4h) which is in turn oxidized to nitrate due to nitrification (Figures 4e and 4f). When the flow paths reach the saturated zone (pressure heads become positive), oxygen contents are decreasing and turnover due to aerobic respiration with associated release and oxidation of ammonium are slowed down (Figures 4e and 4f). Oxygen contents are initially fluctuating in the saturated zone because of pressure head variations (i.e., water table fluctuations due to rain events) which are coupled to the oxygen boundary condition as shown in Figure 2. Once the flow paths reach a depth below the water table of about 0.25 m (pressure heads > = 0.25 m) oxygen becomes limiting and is completely depleted after ∼90 days for the deep and after ∼100 days for the shallow flow path. Under anoxic conditions, increasing concentrations of reduced species (like e.g., iron(II) or sulfide) indicates that the system sequentially shifts to de-nitrification, iron(III)- and sulfate reduction (Figures 4g and 4h). After 250 days, the deep flow path is in a completely reduced state where all oxidized species are depleted (Figures 4e and 4g) and conditions remain reduced until the flow path reemerges at the surface and the water exfiltrates. The shallow flow path reaches completely reduced conditions after 200 days, but shortly before exfiltration oxygen becomes available again and oxidation processes are reactivated (Figures 4f and 4h). The reason why re-oxidation only occurs at the end of the shallow flow path is related to the corresponding exfiltration location. The shallow flow path ends in a shallow, water filled depression where ponded water heights are low enough (pressure heads <0.25 m) for atmospheric oxygen to diffuse into the uppermost layers of the peat so that oxygen gets in contact with the upwelling reduced water. In contrast, the deep flow path which exfiltrates into the stream channel, where ponded water depths are too large to allow resupply of oxygen by diffusion; no reoxidation of reduced species is observed.Animations 1 and 2.

3.3. Spatial Patterns of Hot Spots

[15] In the previous paragraph, results of the biogeochemical simulations for two selected subsurface flow paths were presented in the time domain. A representation in space is depicted in Figure 5 and was generated by interpolating local species concentrations and reaction rates from the biogeochemical model for all flow paths into the 3D spatial domain of the flow model. In Figure 5 the results for the process of sulfate reduction in the model with ml = 0.5 m are presented as an example and plotted for the central transect aligned along Y = 5 m (dashed line in Figure 3a). Figures 5c and 5d show simulated sulfate reduction and sulfide oxidation rates whereas Figures 5e and 5f show the corresponding concentrations of the reaction product (sulfide) and educt (sulfate). Flow paths are shown in Figure 5a and the age of subsurface water (residence time in the subsurface since infiltration derived from particle tracking) is depicted in Figure 5b. In the cross section, areas of intensive sulfate reduction (hot spots) are visible as well as areas where sulfate reduction is practically inactive (Figure 5c). The latter areas are mainly associated with zones of upwelling subsurface water that is in a reduced state and depleted of sulfate (Figures 5e and 5f). They are preferentially located below local depressions. For areas of infiltration, preferentially located below local hummocks, hot spots (Figure 5c) for sulfate reduction can develop because the infiltrating water, originating from the oxygenated unsaturated zone, is rich in sulfate which can be reduced when more reducing conditions are encountered at increasing depth (Figure 5e). This general pattern with local reduction hot spots below hummocks and an inhibition or absence of reduction processes below depressions, is also evident for all other redox-sensitive species (e.g., see plots in Figures S1–S7 in theauxiliary material). In comparison, oxidation processes (iron(III)-,aerobic sulfide oxidation) show a reversed pattern, where local hot spots are preferentially generated below depressions where older upwelling water, rich in reduced species, comes in contact with atmospheric oxygen (Figure 5d). In infiltrating areas, oxidation processes are practically inactive as the freshly infiltrated water carries predominantly oxidized species. Contrary to other oxidation processes, nitrification also occurs below hummocks (Figure 6), because here ammonium, released during turnover of organic material due to aerobic respiration, reacts with oxygen to form nitrate. Nevertheless, even for nitrification the areas showing the most intensive turnover rates are preferentially located below depressions where upwelling water rich in ammonium gets in contact with atmospheric oxygen. As mentioned before, the location of oxidation and reduction hot spots in the virtual wetland models are strongly correlated with the surface micro-topography as illustrated inFigure 7. For each of the models surface topography is displayed in plan view. A binary classification into areas with higher relative elevation (red areas) and local depressions (blue areas) is used for the micro-topography models, whereas the surface of the planar reference model is displayed in graduated colors. Directly to the right of the plots showing the model surface, plan views of high process activity (hot spots) are shown in black, evaluated based on the maximum process activity across the vertical extent of the model at each location in the 2D horizontal domain (again for sulfate reduction as an example). A strong spatial correlation between hot spots and wetland topography can be seen for the micro-topography models. In close proximity to the stream channel (X > 18 m) also surface depressions can be zones of infiltration because of the steep hydraulic gradients toward the adjacent stream channel. Under these conditions depressions are no longer characterized by upwelling of reduced groundwater, which suppresses oxidation but in turn fosters reduction processes (plots for the other implemented processes, which are shown in Figures S8–S12 in theauxiliary material).

image

Figure 6. Results of the biogeochemical simulations exemplarily shown for nitrification of the micro-topography scenario with the mean length 0.5 m. (a) Flow paths along which PHREEQC simulations were performed. Results were interpolated into the 2D cross sections. (b) The age distribution in years of subsurface flow derived from backward particle tracking. (c) Process activity of nitrification (kinetic rate in mol/Ls). Concentrations in mol/L for (d) ammonium, (e) nitrate and (f) oxygen.

Download figure to PowerPoint

image

Figure 7. Plan view of the micro-topography and the planar reference models. Micro-topography is depicted in two categories, red for hummocks and blue for hollows. For the planar reference model elevation classes are shown reflecting the linear slope of the surface. Black areas on the right represent areas of preferential sulfate reduction (hot spots) relative to their surroundings. In general hot spots for reduction processes preferentially form below hummocks and hot spots for oxidation processes below hollows (as shown in theauxiliary material).

Download figure to PowerPoint

[16] Although the formation of hot spots was generally found in both micro-topography models (Figures 8a and 8b) it is significantly more pronounced in the model with larger mean length of the structures. The main reason for that are the more pronounced shallow flow cells that develop in the model with coarser micro-topography. In the model with finer micro-topography (ml-0.25 m) hot spots are less pronounced (the relative difference in reaction rates between the hot spot and its surrounding area is smaller) and spatially more dispersed. This model represents a transition to the planar reference model (Figure 8c), where almost the entire surface area of the model shows infiltration and upwelling conditions are restricted to the zone between X = 9 m and 16 m. As a result biogeochemical process patterns are more uniform and less patchy. The characteristic patterns of hot spots are not only visible along the main direction of subsurface flow as shown in the transects but also in 3D, which is shown in Figure 9 (plots for the mean length 0.25 m model and the planar reference are shown in Figures S13 and S14 in the auxiliary material).

image

Figure 8. Comparison of simulated sulfate reduction hot spots for (a and b) the two different micro-topography scenarios and (c) the planar reference.

Download figure to PowerPoint

image

Figure 9. Fence plots showing the zones of preferential sulfate reduction for the whole 3D domain of the mean length 0.5 m model (3D plots for the mean length 0.25 m and the planar reference case are shown in the auxiliary material).

Download figure to PowerPoint

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

4.1. Hydrological Controls of Hot Spot Formation

[17] Our model scenarios have demonstrated that the basic mechanism of hot spot formation is the same for both micro-topography models. In young, freshly infiltrated subsurface water, electron acceptors are abundant and anaerobic respiration can proceed as soon as the infiltrating water reaches zone 3 where oxygen resupply is assumed to be negligible (τresupply ≪ τdepletion). After depletion of oxygen, denitrification, iron(III) – and sulfate reduction are sequentially initiated for infiltrating water, triggered by the high availability of electron acceptors. Reductive hot spots are generated below infiltration areas located preferentially underneath hummock structures. Initially, below hummocks, infiltrating water passes the unsaturated zone (zone 1) where resupply of oxygen is assumed to occur instantly (τresupply ≫ τdepletion), here aerobic respiration is the only active process. As water infiltrates deeper (zone 2), resupply of atmospheric oxygen is assumed to significantly slow down (reduced diffusivity) and anaerobic processes are being initiated. Denitrification, iron(II) and sulfate reduction become dominant in the part of the saturated zone (zone 3) where the resupply of oxygen is cut off. On the other hand, upwelling zones, where older and already reduced groundwater rises into superficial layers, are characterized by inactivation of reduction processes. Here denitrification, iron(III) and sulfate reduction are inhibited because electron acceptors are not available. Oxidation processes however, are triggered for upwelling areas because here reduced water gets in contact with atmospheric oxygen, which is supplied to zone 2 by diffusion through the saturated pore space. Upwelling of subsurface water preferentially occurs below local depressions. Whether oxidation hot spots can be generated below a depression or not depends on the amount of surface water stored within the superficial depression. If a surface depression is filled with too much water (depth of surface ponding >0.25 m) diffusive penetration of atmospheric oxygen is inhibited and hence the formation of hot spots for oxidation processes is suppressed. In contrast, very pronounced hot spots for oxidation processes can be found below depressions with upwelling groundwater and low ponding depths. Here the saturated pore space is located within zone 2 where diffusion of atmospheric oxygen still exceeds depletion (as opposed to zone 3) and where oxygen can penetrate into shallow layers where it gets in contact with upwelling water carrying high concentrations of reduced species. How fast turnover of reduced species in oxidation hot spots occurs, depends on the availability of oxygen in zone 2, which is controlled by the amount of surface water being stored in the superficial depression. If surface ponding depths are very low (<0.05 m) availability of oxygen is assumed to be very high within zone 2 resulting in fast turnover of reduced species and very pronounced local oxidation hot spots. With increasing surface ponding (0.05 m–0.25 m) oxygen availability drops rapidly resulting in slower turnover rates and less pronounced oxidation hot spots. The availability of oxygen below depressions is therefore mainly controlled by the dynamics of surface water storage, which was found to be highly variable in space and time in wetlands with a hummocky topography, depending on the climatic boundary conditions [Frei et al., 2010]. During intensive rainfall events, surface storage and runoff generation in wetlands with shallow water table can be controlled by a dynamic fill and spill mechanism [Frei et al., 2010]. Depressions are filled with water due to rising groundwater levels during onset of rainfall. With lasting rainfall, isolated ponded depressions start to interconnect with each other building extended surface flow networks [Frei et al., 2010; Antoine et al., 2009]. These surface flow networks can efficiently drain large fractions of the wetland's surface. At times more than 80% of the generated stream discharge may originate from this type of surface flow [Frei et al., 2010]. During high water table conditions, fast diffusion of atmospheric oxygen into the subsurface system is limited to areas of high elevation (hummocks), which remain unsaturated at the surface. During water table recessions and decreasing surface ponding, diffusion of atmospheric oxygen, below depressions with lower surface ponding, becomes more effective in terms of increasing rates for resupply, which triggers oxidation processes for upwelling conditions. Generally field data on oxygen supply in wetlands, its coupling to water table dynamics and peat properties are scarce [Estop-Aragonés and Blodau, 2012], stressing the importance of virtual modeling studies. A special condition can develop during extended dry periods, where depressions become disconnected from the declining water table. Below these disconnected depressions hydraulic gradients may reverse, switching from upwelling to infiltrating conditions. In turn oxidation hot spots will diminish because resupply of reduced species from upwelling groundwater is disrupted. It is reasonable to assume that during droughts hot spot patterns will become less pronounced and may eventually vanish as the system gradually shifts toward a more homogenous distribution of process activities.

[18] In real wetland systems probably more than one mechanism will be responsible for the formation of biogeochemical hot spots [McClain et al., 2003] and a clear separation of the influence of one specific process is almost impossible under field conditions. The simulations presented here, however, demonstrate that heterogeneous process patterns in hummocky wetlands can be explained by the complex re-distribution of redox sensitive solutes in space as being controlled by micro-topography induced, subsurface transport processes and alternating biogeochemical boundary conditions. Furthermore, the presented concept shows that biogeochemical hot spots can be generated without reference to material heterogeneities which often are hardly observable in horizontally relatively homogenous peat soils [Morris and Waddington, 2011; Holden and Burt, 2003; Reeve et al., 2001, 2006; Clymo, 1984]. Of course the presented concept neglects important aspects of real field conditions. Effects of the wetlands vegetation like e.g., root water uptake and its influence on subsurface flow or the special biogeochemical conditions within the rhizosphere [Crow and Wieder, 2005; Knorr et al., 2008; Wachinger et al., 2000] are not considered as well as the potential effects on dispersion for the availability of electron acceptors and donors. Hydrodynamic dispersion may cause a smearing effect where the boundaries between hot spots and surrounding areas are not as sharp and clearly expressed as in an advectively dominated system, because solutes are also re-distributed along concentration gradients (diffusion) and transversally and longitudinally along the advective flow directions (dispersion). The biogeochemical simulations were performed using 5-day time steps, which was necessary because of computational constraints during the flow modeling (e.g., memory overflow, storage limitations). However, it is known that hydrological events at time scales of hours (e.g., single rainstorm events) can influence the biogeochemical processes within wetlands, as e.g., demonstrated for pulses of N2O emission [Goldberg et al., 2010] or high instantaneous CO2 production [Deppe et al., 2010] after wetting. Dynamics at these time scales, however, were not the main focus of this work and at this point cannot be fully accounted for in the present modeling approach because of computational limitations. Further it is known that organic carbon in wetlands typically consists of a fraction of labile components that can be easily utilized by microorganisms (mostly within shallow layers) and more recalcitrant components (more abundant in deeper layers) [Yavitt and Lang, 1990; Reiche et al., 2010; Moore et al., 2007]. Labile organic carbon is not uniformly available as is assumed in our approach. However, there are two main reasons why we think that our assumption of unlimited carbon supply is nonetheless reasonable. First, labile organic carbon availability is higher in shallow peat layers, in which most of the modeled processes occur, mostly due to inputs from the vegetation and high fermentation activity in the rhizosphere [Knorr et al., 2008; Wachinger et al., 2000; Reiche et al., 2010]. Second, we did not include methanogenesis, for which the supply of electron donors will be the key control, as the ubiquitous CO2 may serve as electron acceptor [Achtnich et al., 1995]. Field observations suggested that if alternative electron acceptors were present, the respective process proceeded, while under methanogenic conditions, respiratory activity slowed down and partly ceased [Beer and Blodau, 2007; Knorr et al., 2009]. Nevertheless, the process rate constant, in this case, depends on the quality of organic matter used and is not universal but substrate specific. The application of the Redfield ratio to simulate release of organic bound nitrogen due to decomposition of organic material in terrestrial ecosystems was probably a weak model assumption. Recent literature reported that C:N:P ratios in terrestrial ecosystems vary depending on vegetation types, but on the global scale average at about 186:13:1 for soil biomass and 60:7:1 for soil microbial biomass [Cleveland and Liptzin, 2007]. In our biogeochemical model we assumed that the majority of organic carbon available to microbes originates from vegetation and fermented plant material processed by microorganisms. The Redfield ratio is, however, narrower than the global average observed for soil biomass (106:16:1 compared to 186:13:1) and nitrogen release would be overestimated by our model. That means that the concentrations of ammonia, rates of nitrification and thus also nitrate pools available for denitrification may also be overestimated. Nevertheless, this should translate into slightly longer phases of nitrification or subsequent denitrification only, thus not fundamentally altering spatial patterns of the model output.

4.2. Comparison With Field Observations

[19] Despite these simplifications, the presented model is capable of reproducing spatial variations in pore water concentrations of redox sensitive solutes in the field (Figure 10). Vertical concentration profiles were measured in pore water from six different locations at the Lehstenbach field site, for an area which is comparable in size to the spatial domain of the flow model (10 m × 20 m) [Goldberg et al., 2010; Knorr et al., 2009]. Simulated maxima in nitrate concentrations are found at a depth of ∼0.1 m and not directly at the surface, which agrees with measured data. The observed shift of nitrate concentration maxima has been explained as a result of plant uptake from the upper layers, as plant cover often leads to rapid depletion of nitrate concentrations [Silvan et al., 2005]. However, our biogeochemical simulations suggest an additional explanation for the increased nitrate concentrations at shallow depth: As shown for the cross sections (Figure 6c) high nitrification rates are limited to a relatively thin layer where turnover of ammonium to nitrate is highest. This layer of higher reactivity is the result of the vertical transport of water, which is being enriched with ammonium as it passes the unsaturated zone. Because nitrification rates under aerobic conditions depend on the local availability of ammonium, higher ammonium concentrations result in higher nitrification rates, which can be found directly above the de-nitrification zone where anaerobic conditions trigger rapid nitrate reduction. Similar findings were reported for different field studies [Regina et al., 1999; Goldberg et al., 2010]. Measured depth profiles as shown in Figure 10 are often used to calculate biogeochemical turnover rates based on a simplified approach treating wetlands as diffusion limited systems where the resupply of dissolved electron acceptors/donors is solely controlled by diffusion [Beer and Blodau, 2007; Clymo and Bryant, 2008]. However, model results show that advective transport can be an important component especially for slightly sloping wetlands with micro-topography and can significantly affect the spatial availability and re-distribution of electron acceptors and donors within the subsurface. Vertical concentration profiles simulated in this study suggest that depth variations in the concentrations of redox sensitive solutes observed in the field are probably the result of a complex interplay between three dimensional advective transport processes and biogeochemical reactions, which are in turn controlled by micro-topography moderated interactions between surface and subsurface flow processes and do not arise from pure diffusion and reactions alone.

image

Figure 10. Observed and simulated variations of depth profiles for redox sensitive species (nitrate, iron(II) and sulfate). Grey areas represent envelopes for predicted depth profiles and the black lines (mean +/− standard deviation) actual field observations taken simultaneously at six different locations for an area which is comparable to the model 20 m × 10 m domain at the field site in the Lehstenbach catchment.

Download figure to PowerPoint

5. Conclusions and Implications

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

[20] At the landscape scale, riparian wetlands are commonly assumed to be zones of enhanced biogeochemical transformations (e.g., denitrification) due to anaerobic conditions and large carbon supplies [Johnston, 1991]. Field studies, however, have shown that biogeochemical conditions within wetlands can be quite diverse where most of the biogeochemical turnover may be accomplished in localized zones of higher reactivity (hot spots) [Paul et al., 2006; Knorr et al., 2009; Knorr and Blodau, 2009; Fenner et al., 2011]. Under field conditions, different processes and mechanisms can lead to the formation of hot spots [McClain et al., 2003] depending on the scale of interest. However, explaining such hot spots solely by the heterogeneous distribution of static, physico-chemical properties of the soil [Reeve et al., 2001; Holden and Burt, 2003] may be too simplistic. Our simulations indicate that biogeochemical hot spots can form even in homogeneous peat soils as a result of a dynamic subsurface flow system with (1)complex surface/subsurface interactions where surface micro-topography induces a subsurface flow field that is characterized by a small-scale zonation of in- and exfiltration and(2)hydrological controls of the biogeochemical boundary conditions that either facilitate or suppress redox processes in ex- and infiltration areas. Hence the occurrence of reactivity hot spots does not need to be associated with static heterogeneities in physico-chemical soil properties a priori. In fact, the formation of biogeochemical hot spots in wetland systems may have the potential to alter the hydrodynamic properties of the peat and therefore, typically observed material heterogeneity may result from processes described in this study. The precipitation of iron oxides e.g., which preferentially occurs at oxidation hot spots, can lead to a reduction of the effective porosity and a lower hydraulic conductivity, providing a negative feedback on oxygen penetration; or in areas of reduction hot spots e.g., iron sulfides may become enriched that could be reoxidized upon more severe drying. Our results offer a new perspective on biogeochemical transformation processes in riparian wetlands that provides a dynamic framework to explain process heterogeneity in wetland soils and variability in process rates over time and space. Future work will have to address the interplay between different static (e.g., soil properties, vegetation patterns) and dynamic controls (e.g., flow, temperature and vegetation dynamics) of spatial and temporal variations in biogeochemical process activities in wetlands. It is clear that a mechanistic understanding of the links between hydrologic dynamics and biogeochemical transformations will be crucial for an assessment of climate change impacts on wetland functions and associated ecosystems services. The work presented here can serve as starting point for such an assessment by providing an explorative, mechanistic modeling framework to investigate potential shifts in hydrological and biogeochemical processes including changes in feedback mechanisms caused by changes in climatic forcing.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

[21] This study was funded by the German Research Foundation (DFG, grant FL 631/6-2). Their financial support is greatly appreciated. We would like to acknowledge the constructive comments from Carolyn Oldham and an anonymous reviewer, which greatly helped to improve the final paper. The authors also thank Rob MacLaren, Young-Jin Park, Andrea Brookfield and Ed Sudicky at the University of Waterloo, Canada for their invaluable help with the ins and outs of the numerical code HydroGeoSphere.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Material and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions and Implications
  8. Acknowledgments
  9. References
  10. Supporting Information

Auxiliary material for this article contains 14 additional figures showing further results from the biogeochemical modeling as described in the paper.

Auxiliary material files may require downloading to a local drive depending on platform, browser, configuration, and size. To open auxiliary materials in a browser, click on the label. To download, Right-click and select “Save Target As…” (PC) or CTRL-click and select “Download Link to Disk” (Mac).

Additional file information is provided in the readme.txt.

FilenameFormatSizeDescription
jgrg976-sup-0001-readme.txtplain text document8Kreadme.txt
jgrg976-sup-0002-fs01.tifTIFF image2890KFigure S1. Results of the biogeochemical simulations shown for the aerobic respiration of the micro-topography scenario with the mean length 0.5 m.
jgrg976-sup-0003-fs02.tifTIFF image2303KFigure S2. Results of the biogeochemical simulations shown for the denitrification of the micro-topography scenario with the mean length 0.5 m.
jgrg976-sup-0004-fs03.tifTIFF image3481KFigure S3. Results of the biogeochemical simulations shown for the iron(III) reduction of the micro-topography scenario with the mean length 0.5 m.
jgrg976-sup-0005-fs04.tifTIFF image3519KFigure S4. Results of the biogeochemical simulations shown for the iron(II) oxidation of the micro-topography scenario with the mean length 0.5 m.
jgrg976-sup-0006-fs05.tifTIFF image3551KFigure S5. Results of the biogeochemical simulations shown for the nitrification of the micro-topography scenario with the mean length 0.5 m.
jgrg976-sup-0007-fs06.tifTIFF image3525KFigure S6. Results of the biogeochemical simulations shown for the aerobic sulfide oxidation of the micro-topography scenario with the mean length 0.5 m.
jgrg976-sup-0008-fs07.tifTIFF image3626KFigure S7. Results of the biogeochemical simulations shown for the anaerobic sulfide oxidation of the micro-topography scenario with the mean length 0.5 m.
jgrg976-sup-0009-fs08.tifTIFF image3029KFigure S8. Top view for the micro-topography scenarios and the planar reference.
jgrg976-sup-0010-fs09.tifTIFF image3027KFigure S9. Top view for the micro-topography scenarios and the planar reference.
jgrg976-sup-0011-fs10.tifTIFF image3095KFigure S10. Top view for the micro-topography scenarios and the planar reference.
jgrg976-sup-0012-fs11.tifTIFF image2972KFigure S11. Top view for the micro-topography scenarios and the planar reference.
jgrg976-sup-0013-fs12.tifTIFF image2966KFigure S12. Top view for the micro-topography scenarios and the planar reference.
jgrg976-sup-0014-fs13.tifTIFF image5445KFigure S13. Fence plots showing the zones of preferential sulfate reduction for the whole 3D domain of the mean length 0.25m model.
jgrg976-sup-0015-fs14.tifTIFF image3601KFigure S14. Fence plots showing the zones of preferential sulfate reduction for the whole 3D domain of the planar reference model.
jgrg976-sup-0016-m01.aviapplication/x-troff-msvideo90534KAnimation 1. The animation is showing how redox conditions are changing in time as the water travels along an isolated subsurface flow path belonging to the deep flow system (lower plot). Yellow segments of the subsurface flow pathline indicate the areas where oxygen is available (unsaturated zone) and blue segments indicate areas where no oxygen is available (saturated zone). The animation for the deep flow pathline stops after 1500 days because redox conditions do not change anymore. Concentrations for oxidized species are normalized to their initial and for reduced species to their final concentrations (upper plots). Flow pathlines shown in the animations are not the same as used for Figure 4.
jgrg976-sup-0017-m02.aviapplication/x-troff-msvideo49995KAnimation 2. The animation is showing how redox conditions are changing in time as the water travels along an isolated subsurface flow path belonging to the shallow flow system (lower plot). Yellow segments of the subsurface flow pathline indicate the areas where oxygen is available (unsaturated zone) and blue segments indicate areas where no oxygen is available (saturated zone). Concentrations for oxidized species are normalized to their initial and for reduced species to their final concentrations (upper plots). Flow pathlines shown in the animations are not the same as used for Figure 4.
jgrg976-sup-0018-t01.txtplain text document0KTab-delimited Table 1.
jgrg976-sup-0019-t02.txtplain text document2KTab-delimited Table 2.
jgrg976-sup-0020-t03.txtplain text document1KTab-delimited Table 3.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.