Integrated effects of bioturbation, warming and sea-level rise on mobility of sulfide and metalloids in sediment porewater of mangrove wetlands

Abstract

Global warming and sea-level rise exert profound impacts on coastal mangrove ecosystems, where widespread benthic crabs change sediment properties and regulate material cycles. How crab bioturbation perturbs the mobilities of bioavailable arsenic (As), antimony (Sb) and sulfide in sediment-water systems and their variability in response to temperature and sea-level rise is still unknown. By combining field monitoring and laboratory experiments, we found that As was mobilized under sulfidic conditions while Sb was mobilized under oxic conditions in mangrove sediments. Crab burrowing greatly enhanced oxidizing conditions, resulting in enhanced Sb mobilization and release but As sequestration by iron/manganese oxides. In control experiments with non-bioturbation, the more sulfidic conditions triggered the contrasting situation of As remobilization and release but Sb precipitation and burial. Moreover, the bioturbated sediments were highly heterogeneous for spatial distributions of labile sulfide, As and Sb as presented by 2-D high-resolution imaging and Moran’s Index (patchy at the <1 cm scale). Warming stimulated stronger burrowing activities, which led to more oxic conditions and further Sb mobilization and As sequestration, whilst sea-level rise did the opposite via suppressing crab burrowing activity. This work highlights that global climate changes have the potential to significantly alter element cycles in coastal mangrove wetlands by regulating benthic bioturbation and redox chemistry.

1. Introduction

Mangroves are highly productive ecosystems that store massive quantities of blue carbon in their belowground biomass (Donato et al., 2011). Mangroves play a key role in maintaining and protecting (sub) tropical marine biodiversity and crucial ecosystem services, including regulating key global biogeochemical cycles and serving as buffers to the effects of climate change (Wang and Gu, 2021). Mangroves thrive at interfaces between land and sea, where key environmental factors such as temperature, tide, river runoff, bioturbation and deposition rates are highly dynamic and sensitive to global change (Breitburg et al., 2018; Pan et al., 2020). Because mangroves are intermittently flooded, they also contain sharp redox gradients that can concentrate and accumulate heavy metal(loid)s commonly associated with sulfides (de Lacerda et al., 2020; de Lacerda et al., 2022). Changes in sea level and temperature have the potential to affect estuarine mangrove ecosystems directly (Gilman et al., 2008), or through the adaptation within mangroves in response to changing environments (Woodroffe et al., 2016). Increasing temperatures have the potential to create positive feedback in the global carbon cycle by promoting photosynthesis (Wang and Gu, 2021). In contrast, sea-level rise may decrease soil respiration and increase carbon storage within intertidal mangroves, providing negative feedback in the carbon cycle (Krauss et al., 2014). The potential effects of global change on the cycling of other chemicals in mangrove sediments is less clear, but could significantly impact ecology and biogeochemical cycling (de Lacerda et al., 2022; Osland et al., 2018).

Fiddler crabs are a dominant macrobenthos species and the principal bioturbating organisms in mangrove ecosystems (Sarker et al., 2021). Their burrowing activities greatly affect material cycles and energy flows (Lee, 2008), which can stimulate mangrove growth (Smith et al., 2009). Crab burrows increase the surface area of sediment exposed to the air (Kristensen et al., 2008). The effects of these burrows on both air and water exchange can impact of redox conditions and nutrients, metals, and sulfur cycling (Ferreira et al., 2007; Liu et al., 2022b; Pan et al., 2019; Stahl et al., 2014; Xiao et al., 2022). Burrowing can increase oxygen penetration into sediments, thereby oxidizing carbon (Guimond et al., 2020; Stieglitz et al., 2013; Xiao et al., 2021) and other sediment constituents. The effects of burrows depend on their density and morphology, which vary sharply over time and space. Seaward areas (more flooded) typically have shallower and smaller burrow networks than drier areas, and crab activity, including burrow excavation, is lowest in the winter (Egawa et al., 2021). Both water levels and temperature changes are major components of climate change, and can deeply affect crab distributions and bioturbation extent (Wasson et al., 2019; Wilson et al., 2022). Despite these potential effects, few studies have examined the effect of crab burrowing on transient sediment redox conditions and biogeochemistry.

Arsenic (As) and antimony (Sb) are highly toxic metalloids which have mutagenic and carcinogenic effects on humans (Filella et al., 2002; Tighe et al., 2005), and are considered priority pollutants by the European Union (CEC, 1976) and the United States Environmental Protection Agency (USEPA, 1979). Chemical similarities between the two metalloids have prompted concerns over their enrichment both naturally and anthropogenically in many environments (Arsic et al., 2018; Johnston et al., 2020; Wilson et al., 2010). The geochemical behavior of As and Sb in sediments is highly dependent on their chemical speciation as well as sediment properties such as organic matter, minerals, sulfides, and redox conditions, which can affect retention mechanisms such as adsorption, complexation, and precipitation (Wilson et al., 2010). In particular, the availability of sulfide has considerable impact on As and Sb behavior in sulfurized sediments, via coprecipitation, adsorption, complexation and thiolation (Arsic et al., 2018; Bennett et al., 2017; O’Day et al., 2004). Recent studies have confirmed that As is released under anoxic conditions whereas Sb is mobilized under oxic conditions, which is likely linked to sulfur (S) redox cycles (Arsic et al., 2018; Johnston et al., 2020; Ye et al., 2020). However, the interactions of sulfide with As and Sb are still unclear in bioturbated mangrove sediments, partially due to the lack of high-resolution studies.

Mangroves and other coastal wetlands are heterogeneous and complex environments where biogeochemical zones often vary sharply (at the millimeter or smaller scales), especially in dynamic surface sediments with redox transition zones (Pan et al., 2020; Widerlund and Davison, 2007). The structure, composition and stability of these redox interfaces are of redox crucial for the health of mangrove ecosystems. Crab burrows impact the expression redox gradients between burrow surfaces, water, and air. Observing the effects of burrowing on geochemistry ideally should capture this heterogeneity. Diffusive gradients in thin films (DGT) is an in situ sampling technique that can image bioavailable As(III/V) and Sb(III/V) and other solutes at the spatial resolutions needed to capture the effects of burrowing with minimal disturbance (Davison and Zhang, 1994). These thin films record abundance of bioavailable components in sediment/porewater at a high spatial resolution (Arsic et al., 2018). The binding layer of DGT acts like a plant/root taking up the metals through diffusion (Zhang et al., 2018), and has been proven as an effective method to assess the bioavailability of metal(loid)s in soil and wetland sediments (Guan et al., 2021; Pan et al., 2022a; Zhang and Davison, 2015). Nevertheless, the research studying redox gradients in crab-bioturbated sediments is scarce, and little is known about how these gradients respond to changing environmental conditions. This research used high-resolution DGT films to examine in detail the chemical behaviors of sulfide, As and Sb in mangrove sediments with contrasting burrow networks for the first time.

We hypothesized that temperature and water level-dependent crab burrowing activities change redox conditions and drive changes in mobility of sulfide, As, and Sb in subtropical mangrove wetland sediments. Our experiments directly probe the effect of two critical variables on bioturbation and biogeochemical cycling: (1) the integrated effects of bioturbation and flooding (sea-level rise) on mobilities of sulfide and toxic metalloids through field studies and indoor experiments in highly bioturbated summer months, and (2) the seasonal variability in bioturbation and sediment biogeochemistry as a direct measure of the effect of temperature. To achieve these goals, two subtropical mangrove wetlands in China (Fig. S1) were studied in the field using two-dimensional (2D) DGT-labile S (i.e., free sulfide) and laterally integrated, one-dimensional (1D) depth profiles of As and Sb, and related sediment properties in three seasons (winter, spring, and summer). Based on the field observations regarding ubiquitous crab burrows and extensive bioturbated sediments in summer, indoor incubation experiments were also systematically conducted in the absence of crabs to simulate crab-free, non-bioturbated conditions (a negative control). Moreover, redox gradients of these controls were compared to identical sediments that contained burrows using high-resolution imaging of 2D DGT (for S, As, Sb and Fe). This work provides new insights into the control of bioturbation on redox geochemistry of S, and its effect on soluble (bioavailable) As and Sb, and their response to global environmental changes in estuarine mangrove wetlands.

2. Material and methods

2.1. Study areas

The study areas are located in Jiulong River Estuary (JE) and Tongan Bay (TB) surrounding Xiamen Bay (Fig. S1a), which are typical semi-enclosed bays in southeastern China. This region has suffered from severe pollution of heavy metal(loid)s due to dense industrial activities (Pan et al., 2022b; Wang and Wang, 2017). This region has a regular semidiurnal tide with a tidal range of 3–6 m and a subtropical oceanic monsoon climate with annual rainfall of ~1500 mm, most of which occurs between April and September (http://en.weather.com.cn/), when temperatures are also high. The hydrodynamic forcing in JE is relatively strong due to the large upstream runoff and frequent ship sailing, and the slope of the tidal flat is steep (~8 %), with ~20 m between the low-tide mark and the edge of the mangrove. Two sampling sites, JE1 and JE2, were arranged in the central part and edge of the mangrove, respectively (Fig. S1b). In contrast, TB experiences calm wave activity and has a more gradual slope. Two sampling sites of TB1 and TB2 were similarly arranged in the central part and edge of the mangrove, respectively (Fig. S1c). The four sites were located between the mid-tide and high-tide zones—the higher the tidal zone, the less flooding time in a tidal cycle. The extents of tidal flooding increased from JE1 to JE2 to TB1 to TB2. Fiddler crabs are common macrobenthos that widely inhabit the mangrove whose activities, including, burrowing showed little difference among two areas but large differences among seasons and with tidal zonation, i.e., more burrows in summer than in autumn (Fig. S1d, e) and in high-tide zones (lower flooding period and frequency) compared to low-tide zones (Fig. S1f).

2.2. Field sampling

The DGT probe was composed of a binding gel, an agarose diffusive gel and a Durapore^® PVDF (Millipore). It was deoxygenated with nitrogen and stored in oxygen-free NaCl solution (0.03 M) prior to use. The ZrO-AgI DGT (mixed binding gels of Zr-oxide and AgI for adsorption of oxyanions (As and Sb) and free sulfide, respectively) device, provided by Easy Sensor Ltd., China, was used for field deployment in January (winter and dry season), April (spring and moist season) and July (summer and rainy season). In addition, the HR ZrO-CA DGT (within a high resolution of 10 μm fine intervals of binding particles for 2D imaging of HMs (Ren et al., 2021) device was used for 2D imaging of S, As, Sb and Fe for comparison of burrow and burrow-free sediments in field study in October (autumn and dry season). The devices were inserted into the sediments at each site during the ebb-tide period. After one day, a hollow PVC tube (diameter of 8 cm) was carefully plugged into the DGT-surrounding sediment for both sediment sampling and DGT retrieval (Fig. S1f). Then, a YSI Professional Plus multiparameter meter was used to measure the temperature, salinity, pH and dissolved oxygen (DO) of the overlying water.

2.3. Mesocosm experiment

After field sampling in July, two sediment cores (30 cm depth) and overlying water were collected from JE and TB using disassembled poly (methyl methacrylate) (PMMA) tubes (with an outer diameter of 8 cm and inner diameter of 7 cm) and buckets (25 L), respectively. To eliminate the impact of crab bioturbation in the surface sediments, the crabs were removed and the burrow openings (approximately 2–3 cm depth) in the two cores were filled with adjacent surface sediments, retaining tortuous residual burrows in the deep layers. Then the cores were assembled in empty tubes (more than 2 m length) and the diurnal tide in the intertidal zone was simulated (Fig. S2). This was realized by using a peristaltic pump drawing overlying water from a bucket to the incubator at a rate of 2 m of water head within 6 h and drawing back to the bucket at the same rate, followed by 12 h exposure without overlying water. More details were described in a previous study (Pan et al., 2020).

After three tidal cycles, the DGT probes were inserted into the two sediment cores for the same deployment time as the field research (24 h). Then the highest water level was constantly maintained, simulating a flooding event. After three days, another two DGT probes were applied in the two cores for 24 h.

2.4. Sample preprocessing

In the lab, the partial sediment samples were freeze-dried at −80 °C, then ground and sieved (160 mesh) for later analyses. After cleaning with deionized water to remove the stained sediments, the DGT binding gel was first scanned to obtain the 2D distribution of DGT-labile S using the computer-imaging densitometry (CID) technique (Ding et al., 2012; Teasdale et al., 1999). Then the gels were cut into strips at 5-mm vertical intervals. The strips were eluted and processed for DGT-labile As and Sb according to Wang et al. (2016). The acidified eluates were measured by ICP-MS (PerkinElmer NexION 2000) for DGT-labile As/Sb. All plasticware items used for metalloid processing were pre-cleaned by acid pickling and then deionized water washing.

For 2D imaging, the DGT binding gel was clipped at a size of 10 cm (depth) × 1 cm (width) as described previously (Gao et al., 2015). The gel was mounted onto glass microscope slides avoiding air bubbles and uneven surfaces for further ablating. Then, an ASI RESOLution-LR-S155 laser microprobe equipped with a Coherent Compex-Pro 193 nm ArF excimer laser was used for laser sampling. The spot size of the laser microprobe was selected in the lattice area of 0.1 mm (horizontal) × 0.1 mm (vertical) with a total of 100,000 data points per gel. The vaporized material at each point was characterized by ICP-MS (PerkinElmer NexION 2000), resulting in high-resolution 2D images of DGT-S As, Sb and Fe. To improve the analytical precision, a major element of the binding gel matrix, i.e., ¹³C, was used as an internal normalization standard following Lehto et al. (2012). All detailed operating and analyses of LA-ICP-MS were well described by Gao et al. (2015) and Ren et al. (2021). 1D-depth profiles of element concentrations are created by averaging the concentrations in the 2D DGT profiles at each depth, with variance at each depth indicating something about the heterogeneity at each depth.

2.5. Chemical analyses

For sediment analyses, the grain size was measured by Mastersizer 3000 (Malvern, UK) using fresh samples. All other geochemical analyses were conducted using sieved samples. The total organic carbon (TOC), total nitrogen and total sulfur were determined by a Vario EL III element analyzer (Elementar, Germany). The amorphous Fe/Mn oxides (ASC-Fe/Mn) were extracted using a single-step procedure involving the ascorbic acid (Rozan et al., 2002). Samples were completely digested by a mixture of hydrogen nitrate and hydrofluoric acid for total Fe/Mn (Fe/Mn-T). A modified BCR (Community Bureau of Reference) sequential extraction procedure was used to separate sediment As and Sb fractions (Rauret et al., 1999). All the eluates for Fe, Mn, As and Sb were acidified and determined by ICP-MS.

For ICP-MS analyses, procedural blanks, replicates and reference materials were processed and examined. Internal standards (⁷²Ge, ¹¹⁸In, ²⁰⁹Be) were employed in the calibration of instrumental drift and sensitivity. To evaluate the accuracy of analytical procedures, reference material samples (Agilent P/N 5183–4682) were repeatedly measured after every 10 or 20 samples, with recoveries of 92–107% for As, Sb, Mn and Fe. Triplicate analyses were performed for all samples, with the relative standard deviation (RSD) of less than 10% for every sample and of 1.8–3.6% on average. In addition, reference material samples (Art-No.15.00–0062) and triplicate analyses were performed every 10 or 20 samples for C, N and S, with recoveries of 99–100% and RSD of less than 1%.

2.6. Data processing

For LA-ICP-MS analyses, calibration standards of targeted elements were established via plotting standardized LA signals divided by signals of internal standards versus the corresponding elements mass. Then elements data of the laser were transformed into mass per area ($M_{LA}$, μg cm⁻²) based on the calibration curves (Gao et al., 2015; Liu et al., 2022a; Ren et al., 2021). The concentrations of DGT-labile chemicals ($C_{\text{DGT}}$) were then calculated based on Eq. (1) (Zhang et al., 1995):

$$C_{\text{DGT}}=\frac{M_{LA}\text{Δ}g}{Dt}=\frac{M\text{Δ}g}{DAt}$$ (1)

where $\Delta g$ refers to the thickness of the diffusive layer (0.8 mm); $D$ is the temperature-dependent diffusion coefficient (cm² s⁻¹) (Wang et al., 2016); $t$ is the deployment time (24 h); $A$ and $M$ (μg) is the area of the one-dimensional gel strip (3 mm vertically in this study) and the corresponding mass of element in the eluate calculated as per Eq. (2):

$$M=\frac{C_{\text{e}}V}{f}$$ (2)

where $C_{\text{e}}$ is the ICP-MS measured element concentration in the elute; $V$ is the elute volume; and $f$ is the elution efficiency (Wang et al., 2017). The DGT-As and DGT-Sb concentrations in the vicinity of the sediment-water interface (2 cm below) were used to calculate the diffusion flux between the sediment and overlying water. The flux was calculated based on Fick’s first law, considering the degree of sediment tortuosity (Boudreau,1996).

2.7. Statistical analyses

Measured data were plotted using Origin ver. 2019b software. Other figures were plotted using Corel DRAW ver. 2018 software. The significant differences between different samples were analyzed using SPSS ver. 20.0 software with One-Way ANOVA (followed by Duncan’s multiple range tests) module. The significance level was set at p ≤ 0.05 for the statistical analyses. Linear correlations and multiple regression analyses between different parameters were analyzed and plotted using Pearson correlation and asymptotic analysis within Origin ver. 2019b. Patchiness effect or autocorrelation in 2-D dataset was explored using spatial correlograms of the Moran’s Index (I), computed with R (package “spdep”) following Bivand and Wong (2018) and Thibault de Chanvalon et al. (2015). Scale variance analysis (SVA) was employed to decompose the total variance of dataset to identify the contribution of each scale to the variance following Thibault de Chanvalon et al. (2022).

3. Results

3.1. Solid sediment properties

The overlying water (during the ebb-tide period) and sediment properties are shown in Table S1 for the two field study sites. From January (winter) to April (spring), and to July (summer), the overlying water temperature increased (from 17.2 °C to 21.5 °C, and to 28.5 °C) and DO decreased over the same period. Salinity was more consistent, with a notable decline only in July (6.0–0.70 in JE and 21.6–13.7 in TB), when the monsoon caused riverine flooding. Correspondingly, sediment properties such as Fe/Mn components and TOC were relatively stable, but slightly higher in July, due to either enhanced terrigenous input or sample variability in sediment composition. Total sulfur contents decreased from January to July. Given that sulfide minerals undergo redox cycling with dissolved sulfate in porewater, sulfides are oxidized effectively in summer and accumulate from sulfidization in other seasons. Total nitrogen contents and grain size showed little change.

Although sediment composition was relatively constant over time, there were clear seasonal trends in the association of trace metals with different mineral phases as determined by sequential extractions. The modified BCR sequential extraction procedure separates sediment As and Sb fractions into carbonate and exchangeable (F1, extracted by acetic acid), reducible (F2, extracted by hydroxylammonium chloride), oxidizable (F3, extracted by hydrogen peroxide) and residual (F4, extracted with nitric and hydrofluoric acid fusion) (Rauret et al., 1999). The sum of each fraction is an estimate of As/Sb-T. The residual F4 dominated speciation for both As and Sb, accounting for 86% and 65% of the total content, respectively (Fig. S3). The labile fraction of Sb was highest in July, when both reducible and oxidizable fraction Sb concentrations were highest. The distribution of major labile As fractions (e. g., reducible) followed that of Fe in extractions; consistent with the well-known sequestration of As within ASC-Fe/Mn (amorphous Fe/Mn oxides) (Gorny et al., 2015). Overall, most parameters (except for salinity) showed little differences among the four sites in the same season, indicating similar depositional settings. Interestingly, total labile As/Sb concentrations also increased somewhat in summer. This could reflect increased burial of sediments (that contain Fe oxides), but also could reflect scavenging from the overlying water column in summer months.

3.2. Spatiotemporal distributions of DGT-labile species in the field

In mangroves, one of the most important sediment redox processes is sulfate reduction and cycling. The product of this reduction, sulfide, accumulates in DGT films (DGT-S), which serves as a sensitive redox indicator. DGT-sulfide was generally negligible in the uppermost layers and rapidly increased at a depth corresponding to the transition from the oxidized zone to the zone of sulfate reduction (Rozan et al., 2002). The oxidizing depth decreased (Fig. 1) and hotspot values of DGT-S in anoxic layers increased (Fig. S4) in sequence from JE1 to JE2 to TB1 to TB2 in each season. However, the distribution of DGT-S was so heterogeneous that both hotspots and minima are distributed throughout the profile. Although redox conditions were overall reducing (defined as containing measurable sulfide) at most sites, there was an unexpectedly extensive and deep oxidizing zone (up to ~10 cm depth) observed in the summer (July in JE1&2 and TB1&2; Fig. 1), below which the DGT-S levels were very high.

Fig. 1.

Two-dimensional heatmaps of DGT-labile S with a spatial resolution of 40 μm ×40 μm and depth profiles of DGT-labile As and Sb concentrations in sediments at the four study sites (Jiulong River Estuary (JE1&2) and Tongan Bay (TB1&2)) in January (winter), April (spring) and July (summer), respectively.

DGT-As reflects dissolved As levels in pore water. The vertical trend of DGT-As was somewhat similar to that of DGT-S, with low values near the surface and then variably increased with depth (Fig. 1). DGT-As concentrations in July were much higher at depth in summer (10–15 μg L⁻¹) than in all other seasons (<5 μg L⁻¹), with a sharp boundary corresponding to where DGT-S levels increased sharply. In contrast, the DGT-Sb concentration in January and April was highest in more oxidized surface layers, then decreased with depth. In July, however, it remained stable in most profiles and was much higher than the levels in previous seasons.

3.3. Correlations and cluster analyses for sediment properties and DGT species

DGT-As and Sb were associated with DGT-S levels in each vertical profile (Table S2). DGT-As was significantly positively correlated with DGT-S, while DGT-Sb was inversely correlated with DGT-S. The correlations between DGT-As and DGT-S were weakest in April, when DGT-Sb and DGT-S correlations were strongest. The correlations and cluster analyses for sediment properties and profile-averaged DGT species (Fig. 2) revealed that significant positive correlations and close relationships existed between DGT-Sb, sediment Sb fractions, TOC and ASC-Mn, both of which were decoupled with total sulfur and DGT-S. A significant positive correlation and close relationship existed between ASC-Fe and the predominant mobile As fraction (i.e., reducible F2), indicating the crucial regulation of amorphous Fe oxides on As sequestration.

Fig. 2.

The heatmap correlations and dual hierarchical cluster analyses for sediment and DGT (profile-averaged) properties at four sites in three seasons. The clustered lines represent a family of parameters with close relationships, and the color blocks represent Pearson’s correlation coefficients. Mz denotes mean grain size. TS denotes total sulfur. “**” and “*” denote significance levels at p < 0.01 and p < 0.05, respectively.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.4. Redox zonation in crab-free control mesocosms

Field systems had extensive bioturbation due to crab burrowing. An ideal control for the effect of bioturbation would be a similar field system lacking crabs; however, this is impractical to achieve. To examine redox profiles in the absence of bioturbation, a crab-free control was prepared in the laboratory using intact cores of the same sediments. These non-bioturbated control mesocosms (NB1 and NB2 in Figs. 3 and S5) had much higher DGT-S concentrations than bioturbated field sediments (Jul in Fig. 1), particularly at shallow depths. Interestingly, although controls were not subject to active bioturbation, traces of residual crab burrows are visible at >2 cm depth in controls. The infilled sediment in these burrows were still anoxic but contained lower DGT-S levels than surrounding sediments (Fig. 4). We interpret this as a clear indication that natural infilling of residual burrows effectively inhibited exposure to the air and overlying oxic water, which has been verified as biogeochemical hotspots for Fe and sulfide (Thibault de Chanvalon et al., 2015, 2017). Integrated DGT-S concentrations in these controls were 3.0-fold and 6.1-fold higher than bioturbated field sites in July (Fig 5a, d). DGT-As followed similar trends to DGT-S, with integrated DGT-As levels 5.0-fold and 9.6-fold higher for the whole profile and surficial (0–2 cm) layer, respectively, than their corresponding summer sediment inventories (Fig. 5b, e). DGT-Sb was inversely correlated to DGT-S, with inventories in controls that are 6.4-fold and 4.6-fold lower in crab free controls than in the whole profile and surface (0–2 cm) sediments of bioturbated sediments, respectively (Fig. 5c, f).

Fig. 3.

Two-dimensional heatmaps of DGT-labile S (upper) and depth profiles of DGT-labile As and Sb (lower) in sediments from the non-bioturbation treatments (NB1 and NB2) to flooding treatments (F1 and F2) during the indoor incubation experiment in summer. Sediment cores of NB1/F1 and NB2/F2 were collected from Jiulong River Estuary (JE) and Tongan Bay (TB), respectively. The black dotted box denotes 0–2 cm depth range where crab activities were most intense (Kristensen and Alongi, 2006) and chemicals may change significantly.

Fig. 4.

Three-dimensional images of DGT-labile S showing the trail of residual crab burrows in non-bioturbation treatments of NB1 (a) and NB2 (b) in Fig. 3.

Fig. 5.

Box-plot of DGT-labile S, As and Sb in each whole (0–14 cm) profile (a, b and c) and surficial (0–2 cm) profile (d, e and f) of two mangrove sites (JE1 and TB1) from field burrow (light grey, data from Fig. 1) in July to indoor burrow-free treatment (yellow, data from Fig. 3) to flooding treatment (blue-green, data from Fig. 3) in summer. The oxic depth was indicated by the depth of overall negligible DGT-labile S (purple lines with triangles, data from Figs. 1 and 3). The whiskers represent the variation range in each profile, while the straight line and starlike represents the median and mean value, respectively. The vertical resolution of DGT-labile S was converted to the same interval (5 mm) as As and Sb for mutual comparisons. Capital letters represent significant differences between different sites/treatments at P < 0.05 (One-Way ANOVA, followed by Duncan’s multiple range tests).(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Flooding impedes burrow development while preventing subaerial exposure to oxygen. Flooding increased DGT-S concentrations considerably, particularly in the surface sediments (1.9-fold and 2.5-fold increase relative to even the non-bioturbated controls, Fig. 5a, d). These sediments also had residual burrows, but these burrows did not stand out in DGT-S images as prominently as in non-bioturbated controls (Fig. 3). This change translated to modest increases in median and integrated DGT-As concentrations and decreases in DGT-Sb concentrations (1.5-fold and 2.4-fold increases for As overall and at the surface, 2.1-fold and 2.6-fold decreases in integrated Sb overall and at the surface, respectively) (Fig. 5b, c, e, f). However, in flooding and non-bioturbated controls, their diffusion fluxes were comparable for both As (6.5 and 7.2 μg m⁻¹ d⁻¹, respectively) and Sb (−0.065 and −0.046 μg m⁻¹ d⁻¹, respectively, Fig. S6).

Fig. S7 shows Moran’s Index correlograms and SVA applied in 2-D dataset of non-bioturbation (NB1 and NB2) and flooding (F1 and F2) controls. The results of Moran’s Index showed that the spatial organization of DGT-S was always patchy at the <cm (0.008 to 0.5 cm) scale (I > 0.6, p value <0.001). For farther neighbours (1 cm) the Moran’s index values drop, tending to random organization (Thibault de Chanvalon et al., 2015). The SVA results presented in Fig. S7 (black diamonds) show that most of the variance comes from the scales between 0.5 cm and 1 cm (accounting for more than 60%), then nearly exponential decreased from 0.5 cm to 0.008 cm sacles.

3.5. Effect of season on metal(loid) distribution profiles of bioturbated field sites

In autumn when crab bioturbation begins to diminish, the distribution of DGT-labile species (S, As, Sb and Fe) changed markedly, with contrasting patterns between bioturbated burrow sediment and crab-free controls (Fig. 6a, b). In field sediments with burrowing (Fig. 6b), the chemocline increased to >8 cm in autumn, below which DGT-S and As rapidly increased. DGT-Fe concentrations (Pan et al., 2022b) were quite heterogeneous, with variable but generally low values in surficial environments and in infilled sediments inside burrow traces, and generally higher values at depths >5 cm in intact (burrow-free) sediment. DGT-Sb was essentially inversely correlated to DGT-Fe in both vertical and horizontal dimensions (Fig. 6b, c). In addition, for both burrow sediment and burrow-free sediment, the vertical distribution of DGT-S was in accordance with DGT-As, while DGT-Fe and Sb behaved contrasting, as shown in Fig. 6c and verified by fitting relationship analyses (Fig. S9).

Fig. 6.

Two-dimensional heatmaps of DGT-labile S, As, Sb and Fe in sediments from the JE1 field site in October (autumn) without burrows (a) and containing burrows (b) and comparison of their one-dimensional integrated depth profiles (c). Heatmaps have a spatial resolution of 0.1 mm × 0.1 mm. The black dotted box denotes the 0–2 cm depth range where crab activities were most intense (Kristensen and Alongi, 2006) and thus most likely to impact solute concentrations.

In crab-free controls, the chemocline shallowed in autumn, beginning at ~1 cm, below which DGT-S, As and Fe each increased to their maxima at 2–3 cm depth, respectively, and then fluctuated or decreased somewhat below that depth (Fig. 6c). DGT-Sb was relative enriched in surficial layer (0–2 cm), then decreased constantly with depth, inversely correlated to DGT-Fe. Limited heterogeneity laterally (at constant depth) is present, but its scale was much smaller than in field samples and relative to variation with depth.

4. Discussion

4.1. As and Sb mobility associated with sulfide and seasonal bioturbation

In the field study from January to July, the results showed that DGT-As and S presented consistent distributions and significant positive correlations in nearly all vertical profiles (Table S2). In addition, the beginning depth of elevation and high-value zones (e.g., TB2 in January, JE2, TB1 and TB2 in July) were highly coincident (Fig. 1). This indicates that the existence and development of dissolved sulfide in sediment cores favors As mobilization from solid phases to porewater. There are three principal reasons that aqueous As and sulfide concentrations are highly correlated. First, sediment As is associated with Fe/Mn oxides throughout the sediment profile. Under sulfidic conditions, these sorbents can be reduced by dissolved sulfide, resulting in As mobilization (Burton et al., 2014; Saalfield and Bostick, 2009). Iron sulfide may precipitate, arsenic sulfides are more soluble under typical conditions (Bostick and Fendorf, 2003), e.g., alkaline conditions in this study (Table S1) (O’Day et al., 2004). Second, dissolved As concentrations can increase due to a change in their aqueous complexation, particularly through the formation of thiolated arsenate and arsenite species (Wang et al., 2020). These complexes form most appreciably when free sulfide levels are high (Ye et al., 2020). It may explain why S/As ratios were greater in summer than in other seasons (Fig. 1), owing to more sulfidic conditions. Importantly, thioarsenate formation causes As concentrations even in when little of the Fe oxides have been reduced. Third, sediment bound As(V) could be reduced to As(III), which is more stable in anaerobic environments, and which is generally less strongly retained by oxide minerals (Dixit and Hering, 2003). Solid speciation and/or aqueous speciation would be necessary to determine which of these mechanisms is most important. Overall, the results demonstrated that sulfidic conditions tended to favor As mobilization from solid phases but un-passivate As into solid sulfides in surficial mangrove sediments.

Antimony also adsorbs strongly to Fe and Mn oxides, and forms sulfide minerals under sulfidic conditions (Filella et al., 2002). Despite this apparent similarity, DGT-Sb concentrations were inversely correlated to DGT-As and DGT-S, with DGT-Sb levels higher in oxic surface sediments than at depth (Fig. 1). We interpret this difference in part to differences in how Sb and As adsorb and react with dissolved sulfide. First, Sb is mobilized at the sediment–water interface through the formation of Sb(V), antimonate (Arsic et al., 2018; Ye et al., 2020), which has a weak affinity for Mn/Fe-oxides (Han and Park, 2020). This Sb(V) forms from organic matter-bound Sb decomposition (Tella and Pokrovski, 2012), and the oxidation of Sb(III) sulfides (e.g., Sb₂S₃, Ye et al., 2020). The close relationship between sediment Sb fractions with TOC and Mn in reactive oxides (Fig. 2) confirms that these phases are essential to Sb retention (Rouwane et al., 2016). Under sulfidic conditions, Sb(V) reduction to Sb(III) and its precipitation within Sb(III) sulfides (or complexation with sulfidized organic matter) strongly retards Sb mobility (Hockmann et al., 2020; Leuz et al., 2006). Given that DGT-Sb and DGT-S data in this study consistently show a clear negative correlation between labile S and Sb (Table S2), the formation of Sb sulfide minerals, rather than dissolved sulfide complexes, should exert control of Sb burial rather than Sb mobilization. This is the first high-resolution observation of S-Sb decoupling using DGT films, confirming the previous study that Sb mobility was decoupled from the Fe cycle and was, therefore, more likely linked to sulfur and/or organic carbon (Arsic et al., 2018). Because Sb(V) is weakly retained by oxides at the surface of the mangrove soil, Sb may be mobilized into the water column rather than accumulated in sediments, resulting in incomplete burial.

Seasonal variations in the distribution of labile species within the sediment profile were large (Fig. 1 and Table 1), potentially even leading to changes in sediment inventories of As and Sb. These changes for DGT species were most apparent in July, which is surprising given that higher temperatures in summer can cause faster microbial respiration and lower redox potential, coincident with enhanced reduction of sulfate (Johnston et al., 2020). These higher respiration rates presumably would lead to the development of transient profiles of dissolved metals and sulfide (Sullivan and Aller, 1996). Instead, these data indicate that bioturbation through extensive and active crab burrowing is a more dominant factor in controlling redox state. Other studies also suggest that bioturbation increases the penetration of oxygen into sediments (Pan et al., 2019; Xiao et al., 2019).

4.2. Comparison of (non-)bioturbation and flooding effects on S–As–Sb mobility in mesocosm experiment

A growing number of studies have reported contrasting behavior of As and Sb in sediments under transient redox conditions, confirming that As is released under anoxic conditions that strongly retain Sb (Arsic et al., 2018; Johnston et al., 2020; Ye et al., 2020). Seasonal changes in bioturbation and temperature imparted a large effect on the redox profile in these sediments, consistent with observations in saltmarsh environments (Guimond et al., 2020). Bioturbation overall lead to an increase in the depth of the chemocline, increasing the zone of Sb mobilization and As burial in more oxidized sediments (Fig. 3). Non-bioturbated controls exhibited enhanced sulfide and As concentrations but subdued Sb concentrations compared with natural, bioturbated field environments in summer, when crab bioturbation is most intense during the year (Guimond et al., 2020). The chemocline depth oscillating considerably in field environments from season to season. This is most significantly the effect of temperature on bioturbation and/or sediment microbiota because water levels do not oscillate seasonally. In this study, the contrasting pattern between DGT-S, As and Sb was also notable, especially regarding the seasonal variabilities in redox conditions and bioturbation. As shown in Fig. 3, bioturbation should prevent arsenic release to overlying water, while enhancing Sb release (Fig. S6). As such, crab bioturbation plays a key role in controlling the exchange fluxes of metalloids in sediment-overlying water systems.

Water level is a master variable in controlling sediment redox status because water impedes exchange with the oxygenated atmosphere. As such, it is perhaps expected that flooding experiments reveal the critical role of tide and flooding on redox conditions and the behavior of redox-sensitive solutes (Pan et al., 2021, 2020; Guimond et al., 2020). The interplay of water level with seasonal sediment delivery and plant activity, however, is more nuanced. In summer with rapid production of plant-derived organic matter, increased flooding of intertidal sediments can favor extremely anoxic sulfidic conditions that even approach the sediment surface (Pan et al., 2020). These data suggested that permanent flooding enhanced As mobilization (1.5-fold) and decreased Sb mobility (2.1-fold) relative to previous simulative intertidal environments (Fig. 5a, b, c), largely by suppressing atmospheric oxygen infiltration. However, the diffusion fluxes of both As and Sb were comparable for flooding and non-bioturbated controls (Fig. S6), suggesting that flooding-induced more sulfidic conditions did not cause additional As release or Sb sequestration. It may be explained by attenuated concentration gradients in the surficial layer (0–2 cm) under more sulfidic conditions (Fig. 3). This flooding simulates the effect of sea-level rise, i.e., turning intertidal wetland into submersed mudflat without bioturbation (at least for crabs). While in the wild, sea-level rise may increase erosion, that dissociates deposited sulfides releasing metals to the water column (de Lacerda et al., 2022). Our findings highlighted the crucial role of water levels in influencing physical processes in sediment that in turn affect the fates of trace metals such as As and Sb.

The chemocline (as indicated by DGT-S concentrations) is often quite shallow in sediments (within a few cm of the sediment-water interface), particularly in the absence of bioturbation (Fig. 3). Patchiness effect or autocorrelation in 2-D dataset was explored using spatial correlograms of the Moran’s Index (I), which showed strong patchiness for direct neighbors in horizontal and vertical directions, with a characteristic length of <1 cm. The fact suggests that the impact of sampling thicknesses (roughly 2 cm for DGT) would not result in major bias (Thibault de Chanvalon et al., 2015). The Moran’s Index and SVA also revealed a heterogeneous distribution of DGT-S at <cm scale. As the visual 3-D trail of residual crab burrows shown in Fig. 4, the results further confirmed that the residual bioturbation of crab burrowing could cause significant and local disturbance at the (sub-)cm scale.

4.3. Effect of temperature on redox zonation

We further compared and explored temperature-dependent (non-) bioturbation effects on S-As-Sb mobilities in the field among different seasons as shown in Fig. 7. Similar to above mesocosm experiment, the increased oxic depths were in line with increased DGT-Sb and decreased DGT-S/As mobility in either burrow sediments or burrow-free sediment. Nevertheless, the response of these chemicals to seasonal temperature were variable between burrow and burrow-free sediment. As shown in Fig. 7d, e, f, for burrow-free sediments (winter, spring and autumn), DGT-S and As increased but DGT-Sb decreased with increasing temperature, owing to less oxic conditions and enhanced microbial sulfate reduction which has been shown to be highly temperature-dependent (van Bodegom and Stams, 1999). A recent study has also verified that As and sulfide mobility may be enhanced, while Sb mobility possibly attenuated, by elevated temperatures due to a warming climate (Johnston et al., 2020). Whilst for burrow sediments (autumn and summer), DGT-S and As decreased (3.1-fold and 5.9-fold) but DGT-Sb increased (3.3-fold) compared with burrow-free sediment in autumn, owing to bioturbation as discussed above. Then DGT-S and As further decreased but DGT-Sb highly increased with increasing temperature in the summer. The explanation was that highest occurrence frequency of crabs and burrow density appeared with higher temperature in summer (Egawa et al., 2021), which exerted enhanced bioturbation and thus more oxic conditions whose effects exceeded warming-fueled anoxic biogeochemical processes (sulfate reduction and resultant As/Sb mobility). This finding highlights the crucial role of warming and resultant burrow activities on promoting oxic conditions in sediment, enhancing Sb mobility, and suppressing As mobility, which was opposite that the observations in non-bioturbated sediments.

Fig. 7.

Box-plot of DGT-labile S (a), As (b) and Sb (c) in vertical profiles of site JE1 in field study with different months (data from Figs. 1 and 6), and average values of DGT-labile S (d), As (e) and Sb (f) versus temperature (T). Burrow (light grey) represents sampling of burrow sediment in July and October, while burrow-free (dark grey) represents ordinary sampling of burrow-free sediment in January, April and October. The vertical resolution of DGT-labile S and As/Sb in October was converted to the same (5 mm) as As/Sb in January, April and July for mutual comparisons. Capital letters represent significant differences between different months/sediment types at P < 0.05 (One-Way ANOVA, followed by Duncan’s multiple range tests). The purple lines with triangles represent the trend of relative oxic depth reflected by DGT-labile S in Figs. 1 and 6. T represents the water temperature during field sampling. The whiskers represent the range of variation in each profile, while the horizontal line and star represent the median and mean value, respectively.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

4.4. Heterogeneity in redox status

All samples exhibited significant variation in sulfide, As, Sb and Fe concentrations in DGT films with depth, and significant variation at specific depths. This heterogeneity in part is tied to relict burrow architecture, vertical burrows infilled with new and reworked sediment that exhibits different chemical processes for redox sensitive elements (e.g., Fe and sulfide; Thibault de Chanvalon et al., 2015, 2017). In addition, we would expect that heterogeneous availability of organic carbon in sediments is also important. The intense sulfide production (Fig. 6a) in autumn might be fueled by abundant labile organic matter in (sub)surficial sediments, and resultant mobilization of As from Fe/Mn-bound solid phases (O’Day et al., 2004) and sequestration of Sb in sulfide minerals and thiolated organic matter (Arsic et al., 2018). Lateral heterogeneity in concentrations indicates that these processes of burial and release are also heterogeneous (e.g., DGT-Fe), and affected by burrow structure. Consistent oxygenation resulted in more consistent bioturbation, and thus more uniform chemical conditions (decreased heterogeneity).

The heterogeneity observed in this study is an essential observation to understanding sediment redox processes. To our knowledge, this heterogeneity has not been previously observed. This experiment is the first time 2D imaging of DGT-Sb has been conducted in field environments, and one of few examining porewater As levels (Fang et al., 2018). This visualization of the microscopic (micron-scale) niches for redox-sensitive chemicals, which are nearly impossible to detect with conventional 1D methods. It reminds us of the critical need for high-resolution sediment studies to further understand sediment biogeochemistry at a fine scale. That said, this technique requires considerable equilibration time, and it is not ideal to capture transient changes in concentration that result from more dynamic environmental perturbations, for example, due to tidal fluctuations.

5. Conclusions

This study explored the integrated effects of bioturbation, warming and sea-level rise on mobilities of sulfide, bioavailable As and Sb in subtropical estuarine mangrove wetlands. By combining non-bioturbated/flooding incubation experiments with seasonal field studies, we confirmed the hypothesis that crab bioturbation can promote oxic conditions and further drive contrasting mobilities of sulfide, As, and Sb. In warming summer with active burrowing bioturbation, the rapid aerobic mineralization was expected to enhance Sb remobilization (Fig. 8a, b), which may be released to overlying water via both advection and diffusion stimulated by bioirrigation. In contrast, As was largely scavenged and incorporated by Fe/Mn oxides, whose intense mobilization only occurred in deep anoxic/sulfurized layers, leading to negligible release risk to the overlying water.

Fig. 8.

Conceptual illustration of the mobility mechanisms for As and Sb with seasonal changes in the field study (Jan–Apr (a) and Jul (b)) and with incubation (Non-Bioturbation (c) and Flooding (d)) in the mesocosm experiment (summer). The black arrows denote geochemical processes related to As and Sb (im)mobilization; The red arrows denote migration paths of mobilized As and Sb to the overlying water, thickness of which denote the relative intensity; The text in orange ovals (e.g., Fe/Mn-As) denote solid phases, while the blue text (e.g., S²⁻and Sb(V)) denote aqueous ionic phases; “Adv.+Diff.” denotes advection and diffusion.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Whilst the situation of non-bioturbation stimulated sulfide accumulation, which favored As remobilization but Sb burial inversely (Fig. 8c). Moreover, the bioturbation effects vary with different environmental settings. With global climate change such as warming and accompanying sea-level rise, redox conditions are likely to become more anoxic due to increased flooding but less bioturbation (Fig. 8d). The predictable impact then was completely opposite, with further enhanced As mobility but suppressed Sb mobility. The sea-level rise may shift the intertidal zone through regulating tidal flooding and continental environments which could counteract the above complex geochemical influences. More research is needed to model future climate change effects by setting different hydrological regimes, bioturbation intensities and anthropogenic disturbances.

Data Availability

Data will be made available on request.

Data available in figshare at http://doi.org/10.6084/m9.figshare.19709752