Decoupling between Water Column Oxygenation and Benthic

Mar 11, 2013 - Estuaries are crucial biogeochemical filters at the land–ocean interface that are strongly impacted by anthropogenic nutrient inputs...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Decoupling between Water Column Oxygenation and Benthic Phosphate Dynamics in a Shallow Eutrophic Estuary Peter Kraal,* Edward D. Burton, Andrew L. Rose, Michael D. Cheetham, Richard T. Bush, and Leigh A. Sullivan Southern Cross GeoScience, Southern Cross University, P.O. Box 157, Lismore, Australia ABSTRACT: Estuaries are crucial biogeochemical filters at the land−ocean interface that are strongly impacted by anthropogenic nutrient inputs. Here, we investigate benthic nitrogen (N) and phosphorus (P) dynamics in relation to physicochemical surface sediment properties and bottom water mixing in the shallow, eutrophic Peel-Harvey Estuary. Our results show the strong dependence of sedimentary P release on Fe and S redox cycling. The estuary contains surface sediments that are strongly reducing and act as net P source, despite physical sediment mixing under an oxygenated water column. This decoupling between water column oxygenation and benthic P dynamics is of great importance to understand the evolution of nutrient dynamics in marine systems in response to increasing nutrient loadings. In addition, the findings show that the relationship between P burial efficiency and bottom water oxygenation depends on local conditions; sediment properties rather than oxygen availability may control benthic P recycling. Overall, our results illustrate the complex response of an estuary to environmental change because of interacting physical and biogeochemical processes.



INTRODUCTION Estuaries are crucial biogeochemical filters at the interface between continents and the oceans. The functioning of these systems has been adversely affected on a global scale by anthropogenic nutrient inputs. While the specific impact depends on local conditions, many estuaries that receive excess nutrients are adversely affected through eutrophication, toxic algal blooms and oxygen stress.1−3 Eutrophication is typically associated with external, anthropogenic sources of the key nutrients nitrogen (N) and phosphorus (P). However, significant amounts of nutrients can also be supplied by internal biogeochemical cycling.4−8 Surface sediments in eutrophic systems become more reducing because of increased organic matter deposition and oxygen consumption, which affects sediment biogeochemistry and nutrient cycling. For instance, anoxic conditions promote denitrification, through which nutrient NO3− is eventually lost as N2.9,10 In the case of P, recycling from sediments is enhanced under anoxic conditions due to the absence of amorphous Fe(III) (oxyhydr)oxides that have a strong affinity for dissolved phosphate, and enhanced microbial phosphate release.11−15 In addition, it has been shown that Fe(III) and Mn(IV) oxides may be directly involved in the adsorption and hydrolysis of organic P compounds, thereby affecting P release from sediments.16−19 Despite the importance of sediment biogeochemistry in controlling benthic nutrient fluxes, estuarine research has often focused on the impact of changing bottom water properties on nutrient cycling. In particular, the role of Fe redox cycling in © 2013 American Chemical Society

benthic P cycling is often inferred from bottom water measurements.20−25 There is a lack of understanding of the complex interactions between physical, chemical and biological processes that govern benthic nutrient exchange in eutrophic estuaries. To further our knowledge on estuarine nutrient cycling, this study has combined detailed characterization of physicochemical sediment properties (grain size, physical reworking, Fe−P−S geochemistry) with measurements of sedimentary sulfate reduction rates and benthic dissolved inorganic nitrogen (DIN) and phosphorus (DIP) fluxes in the shallow (average depth ∼2 m), eutrophic Peel-Harvey Estuary (Western Australia). For decades, the Peel-Harvey Estuary experienced extensive dense blooms of (toxic) macroalgae on a seasonal basis, with detrimental effects on water quality and biodiversity.26,27 The algal blooms were linked to elevated phosphate concentrations that resulted from excess input of phosphate from agricultural sources in combination with redox-dependent internal P recycling in a strongly restricted estuarine system.28,29 In 1994, a large ocean channel was dredged to mitigate these issues by increasing exchange between the estuary and the adjacent Indian Ocean. This large-scale human intervention has enhanced flushing of the estuary with relatively nutrient-poor ocean water, thereby also increasing the salinity and vertical Received: Revised: Accepted: Published: 3114

November 29, 2012 March 10, 2013 March 11, 2013 March 11, 2013 dx.doi.org/10.1021/es304868t | Environ. Sci. Technol. 2013, 47, 3114−3121

Environmental Science & Technology

Article

distribution was analyzed by laser diffraction with a Malvern Mastersizer 2000. The wet sediment sample in the second tube was immediately frozen at −20 °C after the N2 purge. Later, sediment samples were thawed in an anaerobic chamber and two ∼1 g (dry weight) subsamples were taken for sequential chemical Fe and S fractionation, respectively.34,35 The Fe fractionation scheme was chosen because it is especially suited for the analysis of sulfide-rich materials. The sequential Fe extraction separates poorly ordered Fe(II) minerals and Fe(III) (oxyhydr)oxides, organically bound Fe, crystalline Fe oxides, and pyrite-Fe. The sum of these fractions is referred to as highly reactive Fe (HR-Fe).36 The sequential S extraction removed acid-volatile S, elemental S, and pyrite-S. The sum of these fractions represents total reduced inorganic sulfur (TRIS). Total P in the sediment was determined as the sum of P fractions extracted from a third wet sample with the extraction scheme developed by Ruttenberg.37 For both Fe and S fractionation procedures, sample tubes were thoroughly purged with N2 during each change of solution to avoid sample oxidation. For Fe, P, and S extractions, one in five samples was analyzed in triplicate: the average relative error for each extraction step was below 10%, except for organically bound Fe (∼25%). Sulfate Reduction Rates. From a second core, 2 cm diameter subcores were taken and injected with 25 μL of a 3.7 MBq Na235SO4 solution at 1- or 2-cm intervals. These subcores were incubated at room temperature for 1 h, then quickly sectioned into 1- or 2-cm slices that were preserved in cold 10% zinc acetate to precipitate dissolved sulfide and terminate microbial activity. Duplicate subcores were incubated for all sites, and a single subcore from site 2 was sectioned immediately after the Na235SO4 injections to determine background radioactivity. The preserved samples were frozen at −20 °C prior to further processing. Later, samples were washed three times with 10% Zn acetate and centrifuged to remove porewater and unreacted 35SO42−. Chromium-reducible sulfur (CRS) was released as H2S by boiling the samples with a solution of 50 g/L elemental chromium in 6 M HCl in a closed N2 atmosphere. The gas flow was first bubbled through a citric acid solution to limit 35SO42− carry-over.38 Then, H2S was trapped as ZnS by bubbling the gas through 10% zinc acetate traps. The radioactivity of 35SO42− in the first Zn acetate wash (measure of total 35S activity in the sample) and the radioactivity of 35S2− in the ZnS traps were determined by liquid scintillation counting. Counts were corrected by subtracting the measured background radioactivity (average + 1σ) measured in the samples of the control core from site 2. Sulfate reduction rates (SRRs) were calculated using the amount of injected 35SO42− that was reduced and incorporated into the CRS pool during incubation. 210 Pb Analysis. The depth profile of 210Pb activity (half-life 22.3 years) can be used to investigate sedimentation processes in young sediments.39 For this study, 210Pb activity was measured in samples from a separate core from site 6. This core had a total length of 36 cm and was sectioned into 2-cm intervals. The samples were dried for 48 h at 105 °C then ground, after which the activities of 210Pb and other γ-emitting radionuclides were determined using γ spectrometry at the Australian Nuclear Science and Technology Organization (ANSTO), Institute for Environmental Research. Benthic Flux Measurements. Triplicate sediment cores of ∼20 cm with 30−40 cm of overlying water were taken from

mixing of the water column. The scale and frequency of algal blooms have decreased substantially since then. Little is known about present-day benthic nutrient dynamics in the estuary, where sediments represent a significant nutrient reservoir.29 However, recent studies have reported the continued accumulation of strongly reducing surface sediments that are unusually rich in Fe monosulfides, at least on a local scale.30 Because of the central role of Fe redox chemistry in P cycling, it is important to understand the role of such sediments in regulating nutrient availability in shallow, eutrophic estuaries like the Peel-Harvey Estuary. Here, we explore the relationships between depositional environment, sediment biogeochemistry, and benthic nutrient dynamics in the estuary. The results reveal a decoupling between bottom water conditions and P cycling: the estuary harbors strongly reducing, monosulfidic sediments that release phosphate to the water column irrespective of bottom water oxygenation.



MATERIALS AND METHODS Field Sampling. Over a two-week period in February 2012, sediments were collected from various sites in the eutrophic Peel-Harvey Estuary in southwest Australia. In the current study, we present results from three selected sites. Site 2 (32°39′59.28″ S, 115°39′19.14″ E) was located in a navigation channel in the eastern part of the estuary, near the mouths of the Murray River and Serpentine River. Site 6 (32°34′31.75″ S, 115°45′29.44″ E) was situated in a navigation channel in the western part of the estuary. Site 8 (32°35′30.45″ S, 115°42′50.75″ E) was located in the center of the estuary, away from navigation channels. At all sites, water column properties, such as dissolved oxygen and temperature, were recorded at 20-cm intervals using a Hydrolab DataSonde 5. Sediment and Porewater Geochemistry. Sediment cores with ∼20 cm overlying water were collected in polycarbonate tubes (length 65 cm, diameter 10 cm). After inspection for surface integrity, cores were sealed airtight and transported on ice to nearby Curtin University for processing. One core per site was sectioned into 1- or 2-cm intervals within 3 h after sampling. Each sample was divided between two polypropylene tubes that were immediately purged with N2 and capped. The first tube was centrifuged for 20 min at 4500 rpm, after which the supernatant was 0.45-μm filtered and directly used for colorimetric assays: Fe2+ and total Fe were measured with the Ferrozine method31 and sulfide (ΣH2S = H2S, HS−, S2− and aqueous sulfide complexes) was quantified with the methylene blue method.32 Part of the remaining porewater was frozen and later analyzed for NH4+ (NH4+ + NH3, NO2−, NO3−) and DIP (H2PO4−, HPO42−) with a flow-injection autoanalyzer.32 Another aliquot was acidified in 1.7 M HCl and later analyzed for SO42− with ion chromatography. Redox potential (Eh) relative to a standard hydrogen electrode and pH in the centrifuged sediment residue were measured with electrodes, after which the sample was purged with N2 and frozen (−20 °C) prior to further processing. Later, part of this sediment was dried for 48 h at 105 °C and subsampled.33 A subsample was ground and decalcified (via two 1 M HCl rinses for 4 and 12 h, respectively) and total organic carbon (TOC) was measured with a LECO CN analyzer. Another part of the centrifuged sediment residue was treated with H2O2 at 70 °C to remove organic matter and rinsed with 1 M HCl to remove shell fragments, after which particle size 3115

dx.doi.org/10.1021/es304868t | Environ. Sci. Technol. 2013, 47, 3114−3121

Environmental Science & Technology

Article

each site. Cores were capped and submerged in a ∼100 L Perspex incubation chamber that was filled with mixed surface water from the three core sites. The chamber was maintained at 22−24 °C (similar to bottom water temperatures in the shallow estuary at the time of sampling) and covered with a shade cloth to simulate light availability at the shallow depths of the coring sites (1.5−2 m). After submersion, the cores were uncapped and magnetic stir bars were placed in each core, ∼15 cm above the sediment surface. The stir bars were driven by a central rotating magnet. Two incubation experiments were performed, with different mixing rates: one during daytime under relatively slow mixing (∼20 rpm of the central magnet) and one during nighttime with faster mixing (∼60 rpm). These treatments were chosen to represent variable water column mixing, which changes in the Peel-Harvey Estuary depending on wind speed.28 After equilibration for 24 h, the cores were capped, and dissolved oxygen (DO), pH, and Eh were measured at the start and again after 4−5 and 9−10 h under slow mixing conditions. At these times, a 50 mL sample was also extracted and 0.45-μm filtered. A small aliquot was poisoned with mercury(II) chloride. Both this subsample and the remaining untreated sample were frozen. Later, the poisoned subsample was analyzed for total organic carbon (TOC) with a Shimadzu 5050 TOC Analyzer. The untreated sample was used for DIP analysis using the molybdenum blue colorimetric assay40 and NH4+, NO2−, and NO3− analysis using a flow-injection autoanalyzer. After completion of the first incubation experiment, the cores were uncapped and allowed to equilibrate again for 24 h prior to the second incubation, which was conducted in an identical manner. Inorganic N and P concentrations were used to calculate the dissolved inorganic N/P ratio, DIN/DIP: (NH4+ + NO3−)/DIP. Net O2 consumption rates and DIN and DIP exchange across the sediment−water interface were calculated from the concentration changes in the overlying water. In addition, expected diffusive fluxes of NH4+ and DIP were calculated from the measured concentration gradients between the bottom water and the uppermost sediment interval (0−1 cm, 0.5 cm average depth) with Fick’s first law

Jd = Ds × ∂C /∂z

Table 1. Selected Sediment Properties for Three Depth Intervals: 0−10 (n = 10), 10−20 (n = 5), and >20 cm (n = 6 or 10)a parameter water depth (m) DO (%) clay (vol %)

silt (vol%)

sand (vol%)

D50b (μm)

TOC (wt %)

TRISc (μmol g−1)

total P (μmol g−1)

depth interval

site 2

site 6

site 8

0−10 10−20 >20 0−10 10−20 >20 0−10 10−20 >20 0−10 10−20 >20 0−10 10−20 >20 0−10 10−20 >20 0−10 10−20 >20

1.5 141 10 (±2) 9 (±2) 12 (±3) 76 (±7) 64 (±9) 52 (±14) 14 (±6) 27 (±9) 36 (±16) 13 (±2) 23 (±14) 35 (±27) 5.2 (±0.6) 4.5 (±0.5) 3.6 (±1.4) 414 (±98) 479 (±107) 383 (±198) 17 (±2) 14 (±2) 9 (±3)

1.4 74 15 (±3) 12 (±2) 12 (±1) 80 (±3) 80 (±7) 71 (±4) 5 (±1) 8 (±5) 17 (±5) 7 (±1) 9 (±1) 17 (±5) 3.6 (±0.2) 3.2 (±0.4) 2.6 (±0.1) 329 (±42) 443 (±118) 396 (±63) 20 (±1) 11 (±3) 6 (±1)

2.0 96 4 (±1) 5 (±1) 5 (±1) 51 (±8) 55 (±8) 62 (±13) 45 (±8) 40 (±9) 33 (±13) 57 (±9) 48 (±13) 38 (±21) 1.4 (±0.3) 1.5 (±0.3) 1.5 (±0.2) 250 (±125) 377 (±129) 443 (±73) 5 (±2) 5 (±3) 5 (±1)

a

Values represent the mean with standard deviations between parentheses. bD50 is the median particle size. cTRIS is total reduced inorganic sulfur.

between 250 and 414 μmol g−1. The reduced sulfur was mostly in the form of sulfide (FeS and FeS2), with only minor amounts of elemental sulfur (typically 10 cm sediment depth (Figure 3b and c). Channels provide an environment that is more sheltered from bottom currents, where light suspended materials such as silt and organic matter can accumulate. These fine-grained sediments have low hydraulic conductivity, which limits exchange between the bottom water and porewaters. Consequently, anoxic decomposition of the relatively abundant OM results in strong porewater SO42− depletion and accumulation of dissolved nutrients and ΣH2S at depth (Figure 3a and b). Surface (0−10 cm) sediments deposited at site 8 were coarser with a larger sand-sized fraction (Table 1), likely a result of winnowing by bottom currents on these more exposed sediments. The

Figure 4. Concentrations of (a) dissolved O2, (b) NH4+, (c) DIP, and (d) molar DIN/DIP ratio over time in the bottom water overlying surface sediment from sites 2 (circles), 6 (triangles), and 8 (squares) during incubation experiment 1 (slow mixing, left panels) and 2 (fast mixing, right panels). Fine dashed line in panel a indicates control, that is, measurement of dissolved O2 in bulk water in incubation chamber. Dashed line in panel d indicates Redfield N/P ratio of 16/1. Symbols represent the mean and error bars represent standard deviation (n = 3).

relatively little between sites, with an average concentration of 0.4 ± 0.2 μmol L−1 in incubation experiment 1 and 0.5 ± 0.2 μmol L−1 in incubation experiment 2. Nitrite concentrations were below the detection limit (