University of Southern Denmark
Drivers of sulfide intrusion in Zostera muelleri in a moderately affected estuary in south- eastern Australia
Holmer, Marianne; Bennett, William W.; Ferguson, Angus J. P.; Potts, Jamie; Hasler-Sheetal, Harald; Welsh, David T.
Published in: Marine and Freshwater Research
DOI: 10.1071/MF16402
Publication date: 2017
Document version: Accepted manuscript
Citation for pulished version (APA): Holmer, M., Bennett, W. W., Ferguson, A. J. P., Potts, J., Hasler-Sheetal, H., & Welsh, D. T. (2017). Drivers of sulfide intrusion in Zostera muelleri in a moderately affected estuary in south-eastern Australia. Marine and Freshwater Research, 68(11), 2134-2144. https://doi.org/10.1071/MF16402
Go to publication entry in University of Southern Denmark's Research Portal
Terms of use This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply:
• You may download this work for personal use only. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying this open access version If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to [email protected]
Download date: 03. Oct. 2021 1 Drivers of sulfide intrusion in Zostera muelleri in a moderately impacted estuary in south-eastern Australia
2
3
4 Marianne Holmer1*, William W. Bennett2, Angus J. P. Ferguson3, Jaimie Potts3, Harald Hasler-Sheetal1,
5 David T. Welsh2
6
7 1. Department of Biology, University of Southern Denmark, Campusvej 55, DK-5230 Odense M,
8 Denmark
9 2. Environmental Futures Research Institute, Griffith School of Environment, Griffith University Gold
10 Coast campus, Queensland, Australia
11 3. Coastal Waters Unit, Science Division, New South Wales Office of Environment and Heritage,
12 Sydney, NSW 2022, Australia
13 14
15 *corresponding author: [email protected]
16
17
18
1
19 Abstract
20 The seagrass Zostera muelleri is abundant in estuaries in Australia and is under pressure from coastal
21 developments. We studied sulfide intrusion in Z. muelleri along a gradient of anthropogenic impact at five
22 stations in the Wallis Lake estuary, Australia. Results showed differences in sediment biogeochemical
23 conditions, seagrass metrics as well as nutrient content and sulfide intrusion along the gradient from the
24 lower estuary (impacted) to the lagoon (unimpacted). Sulfide intrusion was driven by complex interactions
25 and related to changes in seagrass morphology and sediment biogeochemistry and modified by the exposure
26 to wind and wave actions. The sediments in the lower estuary had high contributions from phytoplanktonic
27 detritus, whereas the organic pools in the lagoon were dominated by seagrass detritus. Despite high
28 concentrations of organic matter, sulfide intrusion was lower at stations dominated by seagrass detritus
29 probably due to lower sulfide pressure from the less labile nature of organic matter. Porewater Diffusive
30 Gradients in Thin-films (DGT) sulfide samplers showed efficient sulfide reoxidation in the rhizosphere, with
31 high sulfur incorporation in the plants from sedimentary sulfides likely due to sulfate uptake from reoxidised
32 sulfide. This is a unique adaptation of Z. muelleri, which allows high productivity in estuarine sediments.
33
34
35
36 Keywords
37 Seagrass morphology, sediment biogeochemistry, nutrient enrichment, sulfide intrusion
2
38 Introduction
39 Seagrasses are abundant in estuaries, where they benefit from good light conditions, nutrient inputs from
40 land, shelter and rapid water circulation (Duarte 2002; Boström et al. 2014). These estuarine seagrass
41 meadows provide many essential ecosystem services (e.g. biodiversity, shelter and food for juvenile fish and
42 shellfish, coastal protection, nutrient retention), however, seagrasses in estuaries are also some of the most
43 vulnerable ecosystems due to increasing anthropogenic pressure, with major losses having been observed
44 Worldwide (Waycott et al. 2009). Seagrass loss in estuaries can be attributed to natural and anthropogenic
45 disturbances, and is often linked to eutrophication and coastal development (Baden et al. 2003; Waycott et
46 al. 2009; Boström et al. 2014). High nutrient concentrations in the water column may enhance the growth of
47 filamentous macroalgae and epiphytes, which can shade the seagrass and induce water-column anoxia,
48 leading to seagrass decline. High organic matter loading of the sediments as a result of eutrophication
49 increases the mineralization rates in the sediments and eventually the sulfide pressure on the plants, which
50 may be detrimental to seagrasses and lead to loss of seagrass meadows (Frederiksen et al. 2006; Mascaro et
51 al. 2009).
52
53 Seagrasses grow in reduced sediments in estuaries, where the oxygen penetration is often low due to high
54 loading of labile organic matter (Glud 2008). Mineralization of organic matter occurs primarily by sulfate
55 reducing bacteria and high pools of sulfides can be expected in the rhizosphere sediments (Hansen et al.
56 2000). Mineralization rates and in particular sulfate reduction rates are driven by the lability of the organic
57 matter, where rates are stimulated by increasing inputs of phytoplanktonic detritus (Holmer et al. 2004).
58 Sulfur stable isotope signals in seagrasses have been used to investigate sulfide intrusion into different plant
59 tissues (leaves, rhizomes and roots) (Oakes and Connolly 2004; Frederiksen et al. 2006). Due to bacterial
60 fractionation of sulfate during sulfate reduction, sedimentary sulfides are isotopically lighter than seawater
61 sulfate, and intrusion of sulfides is measured as lower δ 34S signals in the plant tissues. Intrusion of sulfides
62 into seagrasses is commonly observed (Holmer and Hasler-Sheetal 2014), and previous studies of Z. muelleri
63 (formerly Z. capricorni) have shown lower δ 34S values in leaves compared to seawater sulfate, indicating
64 sulfide intrusion (Oakes and Connolly 2004). Many seagrass species are capable of transporting oxygen to
3
65 the rhizosphere, which can influence the chemical composition of the rhizosphere sediments. Brodersen et al.
66 (2015) showed that Z. muelleri releases oxygen from the entire root surface and from the rhizomes, whereas
67 other seagrasses such as Z. marina only release oxygen from the root tip (Jensen et al. 2005). Hasler-Sheetal
68 and Holmer (2015) found several sulfide detoxification mechanisms in Z. marina, including the reoxidation
69 of sulfide to sulfate in the rhizosphere. Rhizosphere sulfate may be taken up by the plant showing an isotopic
70 signal similar to the sedimentary sulfides, as there is no fractionation during the reoxidation process (Hasler-
71 Sheetal and Holmer 2015). If the reoxidation capacity of the seagrasses is exceeded, due to anoxia in the
72 water column or high sulfide pressure from the sediments for example, sulfide intrusion may reduce growth
73 and increase mortality in seagrasses (Mascaro et al. 2009). Sulfide intrusion in seagrasses is complex and
74 controlled by a range of different factors, including plant morphology and sediment biogeochemistry
75 (Holmer and Kendrick 2013; Holmer and Hasler-Sheetal 2014). The biogeochemistry of iron, including the
76 iron content of the sediments, plays an important role in mitigating sulfide intrusion, as dissolved ferrous
77 iron can precipitate sulfides as iron monosulfides or pyrite in the sediments, thereby reducing the sulfide
78 exposure from the porewaters on the seagrasses (Marbà et al. 2007). Furthermore, the oxidation of the
79 sediments by oxygen released from seagrass roots, particularly during the day, can oxidize dissolved Fe2+ to
80 Fe(III) precipitates, which can then chemically oxidize sulfides to elemental sulfur or other sulfur
81 compounds. Declining Fe2+ concentrations during the day have been found in the rhizosphere of Z.
82 capricorni (now Z. muelleri) by the use of Diffusive Gradients in Thin-films (DGT) techniques (Pagès et al.
83 2012). Recent research has shown that iron content can vary significantly along the west coast of Australia
84 and effect sulfide intrusion in five different seagrasses (Holmer and Kendrick 2013). Hence sulfide intrusion
85 in seagrasses is driven by a range of biogeochemical factors, which are all potentially linked to changes in
86 anthropogenic pressures in estuaries.
87
88 Early detection of eutrophication is critical for management of coastal waters to avoid exceeding thresholds
89 of no return (Duarte et al. 2008). As Wallis Lake is under pressure from run-off from land and coastal
90 development it is important to monitor the seagrass meadows as they are key components of the Wallis Lake
91 estuary. Traditional monitoring of water quality has been questioned (Pérez et al. 2008), whereas seagrass
4
92 bioindicators are now used for monitoring in several countries, and are for instance part of the Water
93 Framework Directive in the European Union (Romero et al. 2007; Borja et al. 2013). Seagrass depth limits
94 are widely used, but in shallow coastal waters such as Wallis Lake, light limitation may not be critical for
95 seagrass distribution and other indicators are needed. Cover of epiphytes, blooms of filamentous algae and
96 increase in tissue nitrogen content and change in δ 15N signal in leaves have been observed in areas with high
97 nutrient inputs and are used as indicators of nutrient enrichment (Herbeck et al. 2014; Roca et al. 2015). In
98 addition, the δ 13C of sediments may be used as an indicator of eutrophication, as seagrass meadows
99 accumulate organic matter such as phytoplanktonic detritus and macroalgae debris due to increased
100 sedimentation compared to unvegetated sediments, a likely situation in Wallis Lake, where increased
101 phytoplankton production may increase inputs of organic matter to the sediments particularly at sheltered
102 sites. Determining the δ 13C signal of the potential sources of sedimentation, therefore, makes it possible to
103 identify the sources of organic matter and their origin in the sediments (Kennedy et al. 2010; Fourqurean et
104 al. 2012). Higher sulfate reduction rates due to increased inputs of organic matter to the sediments may
105 increase the sulfide pressure on the seagrasses in Wallis Lake. Intrusion of sulfides, indicated by low δ 34S or
106 high Fsulfide (percentage of plant sulfur derived from sedimentary sulfides), has been suggested as an early
34 107 warning indicators of sulfide pressure, but δ S and Fsulfide have not yet been fully developed as bioindicators
108 (Holmer and Hasler-Sheetal 2015).
109
110 The main objective of this study was to examine potential sulfide intrusion in Zostera muelleri in a lightly to
111 moderately impacted estuary on the south-eastern coast of Australia. Z. muelleri is confined to the south-
112 eastern region of the Australian mainland and the waters surrounding Tasmania, New Zealand and Papua
113 New Guinea and is a significant primary producer in this region (Kerr and Strother 1990; Udy and Dennison
114 1997; Matheson and Schwarz 2007; Kohlmeier et al. 2014). This species has similar morphometrics and
115 growth requirements to the temperate Z. marina, and thus similarities in sulfide intrusion can be expected
116 (Ferguson et al. 2016). Wallis Lake is, like many other estuaries, under pressure from an increasing
117 population and increasing exploitation of the catchment, which increases the nutrient inputs to the estuary
118 (Fiebig 2010). We explicitly tested two hypotheses on sulfide intrusion in Z. muelleri; that sulfide intrusion
5
119 is driven by a) seagrass morphology and/or b) sediment biogeochemistry. The seagrass meadows were
120 examined along a gradient of anthropogenic impact, from the lower estuary (impacted) to the lagoon
121 (unimpacted) and we focused on seagrass morphology (shoot density, above- and below-ground biomass,
122 shoot size), plant nutrient content (as an indicator of eutrophication, Burkholder et al. 2007) and stable
123 isotopic composition, sulfide intrusion, and sediment biogeochemistry. The gradient is considered to be
124 driven by nutrient loading from land the populated northern lower estuary, stimulating phytoplankton
125 production at the expense of benthic primary production and increasing turbidity in the water column (Fiebig
126 2010).
127
128 Materials and methods
129 Wallis Lake is a large wave-dominated barrier estuary (Roy et al. 2001) located on the mid-north coast of
130 New South Wales, approximately 300 km north of Sydney (Fig. 1). The estuary has a catchment area of 1197
131 km2, which is drained by the Wallamba River, Coolongolook River, Wang Wauk River and Pipers Creek.
132 The estuary has a surface area of 99 km2 and a mean depth of 2.3 m, with a salinity range between 31-33 in
133 the central basin. Water column nutrient concentrations are generally low (0.2-0.3 µM NOx, 0.4-0.9 µM
+ 3 + 3 134 NH4 and 0.02-0.37 µM PO4 ; (0.2-0.3 µM NOx, 0.4-0.9 µM NH4 and 0.02-0.37 µM PO4 ; Smith and
135 Heggie 2003) and there is a gradient of anthropogenic impact from the lower estuary to the lagoon (Fig. 1).
136 Bed shear stress is low within the estuary, driven primarily by small tidal currents (<0.1 m s-1). In contrast,
137 bed shear stress in the lake basin is driven by wind wave exposure and tends to be high in shallow areas with
138 significant northeasterly and southeasterly fetches. Zostera muelleri is the dominant seagrass in the estuary
139 with small to large patches of Ruppia megacarpa, Posidonia australis and Halophila ovalis. The estimated
140 cover of seagrasses is 31.9 km2, covering approximately one third of the estuary (Fiebig 2010). There are
141 significant beds of charophytes in the lagoon, and Wallis Lake is surrounded by mangroves and saltmarshes,
142 which contribute terrestrial organic matter to the estuary. The total N load to Wallis Lake is estimated to be
143 953 kg km-2 yr-1(68 mmol m-2 yr-1, Fiebig 2010) and is low compared to other sites (Port Phillip Bay: 200
144 mmol m-2 yr-1 (Harris 1996); Basin d´Arcachon: 460 mmol m-2 yr-1 (de Wit et al. 2001), Limfjorden: 2700
145 mmol m-2 yr-1 (Møhlenberg 1999)).
6
146
147 Sampling
148 Five study stations with extensive cover of Z. muelleri were chosen along a nutrient enrichment gradient
149 from the lower estuarine reach of a freshwater input (Pipers Creek) to the marine dominated lagoon (Fig. 1).
150 Bed shear stress across this gradient was not measured, however, given the fetches of prevailing storms and
151 water depth at each station, it is likely that the shear stress gradient was lowest at Sta. 1 followed by Sta. 5,
152 Sta. 4 and Sta. 2 and highest at Sta. 3. The water depth varied between 1.2 to 2.1 m (Sta. 1: 1.5 m; Sta. 2- Sta.
153 3: 1.2 m; Sta. 4: 1.6 m; Sta. 5: 2.1 m). At each station five cores (i.d. 14.6 cm) were collected to a depth of 30
154 cm for measurements of shoot density and above- and below-ground biomass. An additional eight small
155 cores (cut-off 50 ml syringes) were collected to measure sediment iron and sulfur content and 4 sediment
156 cores (i.d. 8 cm and length 30 cm) were collected for sediment characteristics (density, porosity, C and N
157 pools and stable C, N, and S isotope signals).
158 Four to five replicate Diffusive Gradients in Thin-films (DGT) porewater samplers containing a silver iodide
159 binding layer, as described by (Robertson et al. 2008), were deployed in the seagrass meadows at each
160 sampling station for two days to collect porewater sulfate and sulfide. Potential carbon sources (e.g.
161 epiphytes, drifting macroalgae) to the seagrass meadow were collected by hand in the area during the
162 sampling.
163
164 Processing of samples
165 Cores for seagrass biomass were sieved and seagrasses were separated into leaves, rhizomes and roots. The
166 length and the width of the second oldest leaf were measured on five shoots and the number of flowering
167 shoots was counted as described by Ferguson et al. (2016). The two youngest leaves, rhizomes and roots
168 from ten shoots were collected for analysis of nutrient content and stable isotopes (C,N,S). The plant
169 materials were dried at 60 °C for 1-2 days. Potential carbon sources were dried as above and analysed for
170 stable isotopes (CNS) and nutrient content (CN).
171
7
172 The sediments for iron and sulfide analysis were adjusted to 5 cm and briefly homogenized before
173 preservation in 0.5 M HCl (1 cm3 sediment and 9 mL HCl) and 1 M zinc acetate (20 cm3 sediment and 20
174 mL zinc acetate), respectively. The samples were stored refrigerated or frozen. Iron was measured on the
175 supernatant after dilution according to (Stookey 1970) and Sørensen (1982). The samples for sediment
176 characteristics were sliced at 2 cm intervals down to 10 cm and at 5 cm intervals down to 15-20 cm,
177 depending on the length of the core. Sediments were processed for density, porosity and dry weight. A
178 subsample of dried sediment was analysed for stable C and N isotopes. The sediments preserved in zinc
179 acetate were distilled twice according to the two-step distillation procedure by Fossing and Jørgensen (1989)
180 using either AgNO3 or zinc acetate as a sulfide trap, and separated into AVS and CRS. The precipitated Ag2S
181 was packed into tin capsules together with vanadium pentoxide and analyzed using EA-IRMS (Delta V
182 Advantage Isotope Ratio MS with Thermo Scientific EA) to obtain δ34S in the AVS and CRS fraction.
183 Standards analyzed were IRMS certified reference material EMA-P1, and an Ag2S internal standard. The
184 precipitated ZnS was analysed according to Cline (1969) to determine the sulfide pools.
185
186 Sulfide DGT samplers
187 DGT samplers for measuring dissolved porewater sulfide and sulfate were prepared as described previously
188 (Robertson et al. 2008) in sediment probe housings purchased from DGT Research Ltd. The samplers are
189 composed of a binding layer, which is a polyacrylamide hydrogel containing precipitated silver iodide,
190 overlain with a diffusive layer, consisting of a polyacrylamide hydrogel of 0.8 mm thickness, to control
191 diffusion of solutes into the sampler. Dissolved sulfide diffuses into the sampler from the sediment porewater
192 and irreversibly binds as a black silver sulfide precipitate. The intensity of the black color is proportional to
193 the concentration of dissolved sulfide in the sediment porewaters, and provides the unique capability of
194 examining the two-dimensional distribution of dissolved sulfide in the seagrass rhizosphere. The sulfide
195 distributions were obtained by scanning the retrieved gels, as described previously (Robertson et al. 2008).
196 Porewater sulfate does not interact with the DGT binding layer, but simply reaches equilibrium. Prior to
197 scanning the binding gels, they were back-equilibrated in a known volume of deionized water to obtain
198 porewater sulfate for isotopic analysis.
8
199
200 Analysis
201 Nutrient content and stable isotopes of C and N were obtained on ground plant and sediment material, and
202 analyzed using EA-IRMS (Delta V Advantage Isotope Ratio MS with Thermo Scientific EA) with
203 acetanilide and urea used as standards (Working standards, IVA- Analysentechnic). The contribution of
204 different primary producers to sediment organic carbon was calculated by using the mass-balance model
205 IsoSource 1.3 with increments of 1% and tolerance of 0.1, based on the primary producer’s δ13C signatures.
206
207 Plant total sulfur (TS) and δ34S was analyzed as described above by EA-IRMS by weighing ground plant
208 material into tin capsules together with vanadium pentoxide. The water samples collected were kept frozen
34 209 until analysis for δ Ssulfate. Samples were prepared by boiling under acidic conditions followed by
34 210 precipitation of sulfate with BaCl2 as BaSO4, and analysed as described for δ S in sediments and plants. The
211 sulfate eluted from the DGT samplers was precipitated and measured as described for the water samples to
212 obtain δ34S for porewater sulfate. It was not possible to measure the collected sulfide on the samplers for δ34S
213 due to instrumental limitations, and they were only used to illustrate sulfide distributions in the rhizosphere
214 sediment.
215
216 The percentage of sulfur in the plants derived from the sediment (Fsulfide) was calculated according to
217 Frederiksen et al. (2006):
(%) = × 100 218 𝟑𝟑𝟑𝟑 𝟑𝟑𝟑𝟑 𝛅𝛅 𝑆𝑆𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 − 𝛅𝛅 𝑆𝑆𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝐹𝐹𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝟑𝟑𝟑𝟑 𝟑𝟑𝟑𝟑 34 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠34𝑠𝑠 34 219 where δ Stissue is the value measured in the leaf,𝛅𝛅 rhizome𝑆𝑆 or− root,𝛅𝛅 𝑆𝑆 and δ Ssulfide and δ Ssulfate are the values
220 measured in sediment and porewater samples, respectively.
221
222 Statistics
223 All figures and statistical analysis were conducted in GraphPad Prism version 6 for Mac (GraphPad
224 Software, La Jolla, CA, USA). Analysis of variance (ANOVA) was used as the test assumptions were
9
225 fulfilled (visual inspection and Brown-Forsythe test). For results of ANOVA that showed significant effects
226 (p<0.05), the Tukey’s honestly significant difference post-hoc procedure was used to determine significantly
227 different means. Linear regressions were conducted when test assumptions were fulfilled and reported with
228 goodness of fit (R2) and p-values (F-tests to test significant deviations from zero). Data were adjusted for
229 outliers using the Grubbs’ test. If not stated differently all mean values are presented with standard error of
230 mean (mean ± SEM).
231
232 Results
233 The water column concentrations of TN decreased by a factor of 3 from Sta. 1 (~350 µg l-1) to the marine
234 lagoon sites (Sta. 4 and Sta. 5 ~100 µg l-1). The sediments had the lowest density (1.2-1.3 gWW cm-3) and
235 highest porosity (0.79-0.97) at Sta. 1 and Sta. 4 and 5, indicating less wind and wave exposed stations
236 compared to Sta. 2 and Sta. 3, which had highest density (1.8-1.9 gWW cm-3) and lowest porosity (0.62-
237 0.66). The sediment organic matter pools represented a combination of expected high organic matter loading
238 and exposure, where Sta. 1 had POC (3.4%DW) and PON pools (0.36%DW) in between the most sheltered
239 (Sta. 4 and Sta. 5) and most exposed sites (Sta. 2 and 3, Fig. 2A and 2B, ANOVA p<0.05). This resulted in a
240 large difference in C:N ratio among stations, with the lowest ratio (C:N = 11-15) at stations 1, 4 and 5 and
241 the highest at Sta. 3 (C:N = 67, Fig. 2C, ANOVA p<0.05). Sediment Fe content was similar among stations
242 (59-74 mmol Fe cm-3; Table 1) and did not show any response to eutrophication or exposure gradient.
243 Similarly the AVS pools showed no response and were quite similar at all stations (0.5-2.7 µmol cm-3) and
244 low compared to CRS. In contrast, the CRS pools varied along the gradient with low pools at Sta. 3
245 (27.2±9.5 µmol cm-3) and the maximum at Sta. 1 (133.0±11.4 µmol cm-3, Fig. 2D, ANOVA p<0.05). The
34 34 246 δ SAVS ranged between -19.8‰ and -23.3‰ and δ SCRS between -23.5‰ and -28.1‰ with no distinct
247 pattern between stations (Table 1). Sulfate was collected by equilibration within the DGT samplers, and the
34 248 porewater concentration in the rhizosphere sediment varied between 19.2-31.6 mM and the δ Ssulfate was
249 20.84‰-21.19‰ (Table 1) with no major differences between stations. Unfortunately it was not possible to
250 collect DGT profiles of porewater sulfides consistently at the 5 stations due to harsh weather conditions.
251 Scanned images of selected DGT samplers from Z. muelleri rhizosphere sediments at Sta. 1 and Sta. 4
10
252 showed what appeared to be an absence of dissolved sulfide around two roots that were probably in contact
253 with the DGT samplers during deployment (Fig. 3).
254
255 Seagrass morphological parameters varied among stations reflecting both the eutrophication gradient and
256 degree of exposure and did thus not show a clear linear relationship with eutrophication or exposure. Shoot
257 density of Z. muelleri varied by a factor of 4 between the five stations with lowest density of 836±335 shoots
258 m-2 at Sta. 2 and highest with 3285±1241 shoots m-2 at Sta. 4 (Fig. 4, ANOVA p<0.05). This was reflected in
259 above-ground biomass, which showed the same pattern, and the difference between Sta. 2 and Sta. 4 was 4.8
260 fold (98±49 gDW m-2 and 487±287 gDW m-2, respectively, ANOVA p<0.05). The ratio of above- to below-
261 ground biomass (A:B) ranged from 1.3 to 1.9 with the highest ratio at Sta. 1 and 5 (Table 2, ANOVA
262 p<0.05). Shoot morphology varied among stations with the longest and heaviest shoots at Sta. 5 (580±130
263 mm; 276±78 mgDW shoot-1, ANOVA p<0.05), whereas there were no significant differences in the
264 parameters at the other 4 stations (299-365 mm, 113-153 mgDW shoot-1; Table 2, ANOVA p<0.05).
265 Intensity of flowering was highly variable among and within stations with flowering at Sta. 1-4 (36-137
266 flowers m-2), but no flowering at Sta. 5 (Table 2).
267
268 The N content in the plant leaves reflected the eutrophication gradient and showed a significant decrease
269 from Sta. 1 to Sta. 5 of 1.54 to 1.29 %DW (ANOVA p<0.05, Table 3). The N content in the rhizomes (0.39-
270 0.75 %DW) and roots (0.66-0.85 %DW) was approximately half of that in the leaves, but did not show the
271 same decreasing trend along the gradient (Table 3). The TS content was highest in the roots at Sta. 1
272 (1.03±0.19 %DW) and lowest at Sta. 3 (0.62±0.12 %DW) (Table 3, ANOVA p<0.05) and showed no clear
273 trend along the eutrophication gradient. The TS content was the same in leaves and rhizomes for Sta. 1 and
274 Sta. 2, whereas TS was highest in the leaves at Sta. 3-5. Fsulfide was highest in the roots (57-80%), decreasing
275 in the rhizomes (38-57%) and lowest in the leaves (18-48%, Table 3). Fsulfide was significantly lower at Sta. 3
276 compared to end points of the gradient (Sta. 1 and Sta. 5), whereas there were few significant differences
277 among the other stations (ANOVA p<0.05). There were positive linear correlations between TS and Fsulfide
278 for each type of tissue (Fig. 5, R2= 0.50-0.72).
11
279
280 The δ13C varied between -9.5‰ to -11.4‰ in the leaves, -9.9‰ to -11.5‰ in the rhizomes and -9.7‰ to -
281 11.4‰ in the roots without any clear pattern among stations (Table 4). The δ15N was highest at Sta. 3 in all
282 tissues (4.3‰-4.9‰) and lowest at Sta. 5 (0.95‰ in leaves increasing to 2.54‰ in roots; Table 2, ANOVA
283 p<0.05). Seagrass leaves had very few epiphytes, though at several stations (Sta. 1, Sta. 4-5) the leaves were
284 covered by particulates; probably resuspended particulates due to high winds during the sampling period. We
285 collected Halophila sp. and some drifting macroalgae at the stations (Table 3) and δ13C and δ15N varied
286 between -14.0‰ to -18.1‰ and 0.1‰-2.7‰, respectively, for these plant tissues. Results from the IsoSource
287 model showed an increasing average contribution of seagrass detritus to the sediment organic carbon pool
288 from 7.5% at Sta. 1 to 71-75% at Sta. 4 and 5 (Table 4).
289
290 There were significant correlations between Fsulfide and seagrass morphology, seagrass nutrient content and
291 sediment biogeochemistry (Table 5). The strongest correlations were with sediment biogeochemistry: a)
292 contribution of seagrass to the organic matter pool (negative correlation R2=0.62; p<0.0001), followed by b)
293 sediment POC (negative correlation, R2=0.48; p=0.0002) and c) sediment PON (negative correlation,
294 R2=0.41; p=0.0007). Seagrass nutrient enrichment, measured as increasing N content in the leaves (R2=0.32;
15 2 295 p=0.0037) and increasing N signal in the leaves (R =0.49; p=0.0001), also correlated with Fsulfide, whereas
296 the correlations with seagrass morphology were more variable (negative correlation with leaf width R2=0.47;
297 p=0.0002, negative correlation with above-ground biomass R2=0.38; p=0.0014, negative correlation with
298 below-ground biomass R2=0.20; p=0.027, negative correlation with shoot density R2=0.21; p=0.024).
299
300 Discussion
301 The Fsulfide in Z. muelleri was high in Wallis Lake (up to 80% in the roots) compared to other seagrass species
302 in Australia (Holmer and Kendrick 2013) and similar to Zostera marina, where intrusion of sulfides can be
303 high with up to 80-90% of the sulfur in the plants derived from sedimentary sulfides (Holmer and Hasler-
304 Sheetal 2014). Our two hypotheses on possible links with a) seagrass morphology and b) sediment
305 biogeochemistry on sulfide intrusion were confirmed, but as both seagrass morphology and sediment
12
306 biogeochemistry changed along the gradient, there were several potential effects of sulfide intrusion on Z.
307 muelleri and several possible drivers of sulfide intrusion into Z. muelleri. The negative correlations between
308 Fsulfide and seagrass above-ground and below-ground biomass and shoot density suggest that sulfide intrusion
309 negatively affected the above- and below-ground biomass and shoot density. This is consistent with findings
310 by Mascaro et al. (2009), where sulfide intrusion stimulated by organic matter additions to the sediments and
311 by hypoxia in the water column during day and night time, resulted in reduced leaf growth rate, increased
312 mortality and lower above- and below-ground biomass in Z. marina. Sulfide toxicity damaging the seagrass
313 tissues, in particular the meristems, were considered the main driver of lower seagrass performance. It is
314 possible that the negative correlations in Wallis Lake could also be unrelated to sulfide toxicity, and instead a
315 result of lower intrusion of sulfides in dense Z. muelleri meadows with high above- and below-ground
316 biomass driven by oxidation of the sediments from oxygen released by the below-ground biomass. Brodersen
317 et al. (2015) and Koren et al. (2015) found high leakage of oxygen from Z. muelleri compared to Z. marina
318 as oxygen was released along the entire root and to some extent from the rhizomes. The radial oxygen loss
319 (ROL) is much higher in Z. muelleri compared to Z. marina, where oxygen is only leaked to the sediments
320 through the root tips (Jensen et al. 2005), suggesting that Z. muelleri has a larger capacity to oxidize the
321 rhizosphere sediments compared to other seagrass species and thereby reduce sulfide intrusion in particular
322 in dense meadows. This hypothesis is supported by previous findings, where major shifts in porewater Fe2+
323 and sulfide concentrations between light and dark conditions in Z. capricorni (now Z. muelleri) rhizosphere
324 sediments demonstrating a significant capacity of Z. muelleri to buffer sulfide accumulation in rhizosphere
325 sediments (Pagès et al. 2012). In our study, two DGT samplers were by coincidence placed close to Z.
326 muelleri roots and they showed that porewater sulfide was absent along the length of the roots (Fig. 5)
327 confirming a high reoxidation of sulfides by leaked oxygen. Furthermore AVS pools were very low and the
328 sedimentary sulfide pool was dominated by the CRS fraction (>96% of TRS). High pools of CRS are
329 generally found in more oxidized sediments due to formation of elemental sulfur and pyrite (Holmer 2009).
330 As Fsulfide in Z. muelleri is high compared to other seagrasses despite the oxidized rhizosphere, it is most
331 likely that a major fraction of the sulfides are reoxidized to sulfate immediately before entering the plant in
332 the rhizosphere around individual roots. Reoxidation of sulfides to sulfate is considered an important
13
333 detoxification mechanism in Z. marina accounting for up to 12% of the sulfide intrusion (Holmer and
334 Hasler-Sheetal 2014), and this mechanism is possibly even higher in Z. muelleri due to the extended
335 oxidation of the rhizosphere. This may represent an important adaptation by Z. muelleri to increase survival
336 in reduced estuarine sediments. The negative correlation between leaf width and sulfide intrusion further
337 suggests higher release of oxygen from the larger seagrasses with wider leaves, a relationship which should
338 be explored further.
339
340 The correlations between Fsulfide and sediment biogeochemistry varied along the gradient. Most surprising
341 was the negative correlations between Fsulfide and pools of sediment organic carbon and nitrogen, showing
342 that the sulfide intrusion decreased as the organic matter pools increased. This relationship was driven by
343 high pools of organic matter in the lagoon sediments, where the organic matter probably accumulated as
344 these sites were the most sheltered along the gradient. The organic matter pools in the lagoon sediments were
345 high (average 6.5% POC and 0.5% PON) compared to global averages (1.5% POC and 0.15% PON) in
346 seagrass sediments (Kennedy et al. 2010), which could be due to increased trapping efficiency in the dense
347 meadows and lower physical energy at these stations. The δ13C sediment analysis showed that the buried
348 organic matter primarily originated from seagrass detritus, as >70% of the sediment organic carbon pool
349 could be attributed to seagrass detritus at these two stations, which is high compared to global average of
350 about 50% in seagrass sediments (Kennedy et al. 2010). As seagrass detritus is more recalcitrant compared
351 to phytoplanktonic detritus the rates of sulfate reduction and thus sulfide production is most likely lower in
352 the lagoon sediments compared to the other stations. In the lower estuary, the contribution of seagrass
353 detritus to the sediment organic matter was much lower and only 7.5% at Sta. 1, and the high contribution of
354 phytoplanktonic detritus was also reflected in lower sediment C:N ratio at Sta. 1 and Sta. 2. We measured
355 higher sediment oxygen uptake at these two sites compared to Sta. 3 (SOU was only measured at Sta. 1-3;
356 Sta. 1: 48.0 mmol m-2d-1; Sta. 2: 32.8 mmol m-2d-1, Sta. 3: 30.0 mmol m-2d-1 (average, n=3) unpublished
357 data) indicative of higher rates of mineralization and supporting higher sulfate reduction rates at Sta. 1 and
358 Sta. 2. The correlations between Fsulfide and sediment sulfide pools were not significant, probably due to
359 interactions between seagrass oxidation capacity (e.g. below-ground biomass) and changes in sediment
14
360 biogeochemistry (e.g. sulfate reduction rates) along the gradient. The range of sediment organic matter
361 content and other sediment characteristics found in Wallis Lake is similar to other studies showing that Z.
362 muelleri has large plasticity to grow in variable environments (Matheson and Schwarz 2007; Kohlmeier et
363 al. 2014).
364
365 Nutrient enrichment of Z. muelleri in Wallis Lake was indicated by higher N content in the leaves in the
366 lower estuary, decreasing towards the lagoon. The N enrichment in the leaves at Sta. 1 and Sta. 2 may, in
367 addition to inputs of N from terrestrial land run-off, derive from remineralized phytoplanktonic detritus in
368 the rhizosphere sediments which was higher at Sta. 1 and Sta. 2 as discussed above. Wallis Lake appears to
369 be a productive estuary, and above-ground and below-ground biomass was high compared to other studies of
370 Z. muelleri in the distribution area (Kerr and Strother 1990 (South-Australia and Tasmania); Mellors et al.
371 2005 (Great Barrier Reef); Matheson and Schwarz 2007 (New Zealand); Kohlmeier et al. 2014 (New
372 Zealand)) and given the high coverage of seagrasses in Wallis Lake, this suggests that seagrass meadows are
373 an important component of the Wallis Lake ecosystem. The seagrass morphology varied significantly among
374 stations, but there was no clear pattern from the lower estuary to the lagoon. The plants in the lagoon had the
375 heaviest, longest and widest leaves. Shoot density at Sta. 5 was only one third of Sta. 4 suggesting light
376 limitation due to self-shading, as the leaves were also longest at this station. Sta. 5 was 0.5 m deeper than
377 Sta. 4 and with fine grained sediment, which probably negatively affected the light climate during
378 resuspension events as experienced during sampling. Typical responses of Z. muelleri to decreased light
379 conditions are broadening and elongation of leaves (Abal et al. 1994; Peralta et al. 2000). In the lower
380 estuary the shoots were smaller and the meadow less dense and the total biomass was lower, which could be
381 a result of several factors including light climate, physical energy and sulfide toxicity. There was no distinct
382 pattern in the above:below ground (A:B) ratio along the gradient, probably due to this mix of factors acting
383 on the ratio (Ferguson et al. 2016). The ratios were similar to other studies, but the range was lower,
384 suggesting less variation in the environmental conditions compared to other study sites (Matheson and
385 Schwarz 2007; Kohlmeier et al. 2014). There were no ecosystem signs of eutrophication in the seagrass
386 meadows as the plants had few epiphytes and no blooms of filamentous algae were observed at the sampling
15
387 stations. The N content of the leaves thus appears to be a more sensitive indicator of nutrient enrichment.
15 388 The positive correlations between Fsulfide and N content in the leaves and between Fsulfide and δ N signal in
389 the leaves suggests increasing intrusion of sulfides towards the source of nutrients and thereby the first signs
390 of the potential negative impact of nutrient enrichment on seagrasses. Monitoring of N content and δ 15N
391 signal of leaves along with Fsulfide in the seagrasses are potential candidates as early indicators of nutrient
392 enrichment and sulfide pressure on the seagrasses, which is consistent with findings in Z. marina in the
393 Black Sea (Holmer et al. 2016). It should be further explored if these early indicators can supplement the
394 traditional monitoring of seagrass meadows, which typically consists of monitoring of water quality,
395 sediment parameters and seagrass distribution. Seagrass distributions show large seasonal and inter-annual
396 variations making the interpretation of monitoring data difficult. In contrast, early warning indicators of
15 397 seagrass stress, such as N content, δ N signal of leaves, and Fsulfide, could provide a more reliable approach
398 to seagrass monitoring.
399
400 In conclusion, Z. muelleri in Wallis Lake grows in dense meadows from the lower estuary to the lagoon
401 despite nutrient enrichment in the lower estuary. There were signs of nutrient enrichment, as the N content of
402 the leaves was enriched in the lower estuary and at the same time the sediments had higher contribution of
403 organic matter derived from phytoplanktonic detritus. The sulfide intrusion was generally high in Z. muelleri
404 and linked to N enrichment. As higher sulfide intrusion was concurrent with lower shoot density and
405 biomass, the sulfide intrusion may be an early sign of the negative effects of nutrient enrichment. The
406 deployment of DGT samplers, however, provided evidence of high oxidation of sulfides in Z. muelleri
407 rhizosphere sediments, which suggests that most of the sulfur entering the plants was in reoxidized and non-
408 toxic forms. High oxygen release from Z. muelleri provides high tolerance for sulfide intrusion and is a
409 unique adaptation, which allows high abundance and productivity of Z. muelleri in estuarine sediments.
410
411 Acknowledgement
412 The authors would like to thank the team of Angus Ferguson (Peter Davies, Aaron Wright, Julia Fortune,
413 Michael Sutherland, Chris Baiada), Ryan Stewart and the technicians in the Ecolab, SDU for their assistance
16
414 with the analysis. The study was supported by a collaboration grant between SDU and Griffith University.
415 MH was supported by FNU grant no. 12-127012.
416
417
17
418 References 419 420 Abal, E.G., Loneragan, N., Bowen, P., Perry, C.J., Udy, J.W., and Dennison, W.C. (1994) Physiological and 421 morphological responses of the seagrass Zostera capricorni Aschers, to light intensity. Journal of 422 Experimental Marine Biology and Ecology 178(1), 113-129.
423 424 Baden, S., Gullström, M., Lundén, B., Pihl, L., and Rosenberg, R. (2003) Vanishing Seagrass (Zostera 425 marina, L.) in Swedish Coastal Waters. AMBIO: A Journal of the Human Environment 32(5), 374-377.
426 427 Borja, A., Elliott, M., Andersen, J.H., Cardoso, A.C., Carstensen, J., Ferreira, J.G., Heiskanen, A.-S., 428 Marques, J.C., Neto, J.M., Teixeira, H., Uusitalo, L., Uyarra, M.C., and Zampoukas, N. (2013) Good 429 Environmental Status of marine ecosystems: What is it and how do we know when we have attained it? 430 Marine Pollution Bulletin 76(1–2), 16-27.
431 432 Boström, C., Baden, S., Bockelmann, A.-C., Dromph, K., Fredriksen, S., Gustafsson, C., Krause-Jensen, D., 433 Möller, T., Nielsen, S.L., Olesen, B., Olsen, J., Pihl, L., and Rinde, E. (2014) Distribution, structure and 434 function of Nordic eelgrass (Zostera marina) ecosystems: implications for coastal management and 435 conservation. Aquatic Conservation: Marine and Freshwater Ecosystems 24(3), 410-434.
436 437 Brodersen, K.E., Nielsen, D.A., Ralph, P.J., and Kühl, M. (2015) Oxic microshield and local pH 438 enhancement protects Zostera muelleri from sediment derived hydrogen sulphide. New Phytologist 205(3), 439 1264-1276.
440 Burkholder, J.M., Tomasko, D.A., and Touchette, B.W. (2007). Seagrasses and eutrophication. Journal of 441 Experimental Marine Biology and Ecology 350(1-2), 46-72.
442 443 Cline, J.D. (1969) Spectrophotomitric determination of hydrogen sulfide in natural waters. Limnology and 444 Oceanography 14(3), 454-458.
445 446 de Wit, R., Stal, L.J., Lomstein, B.A., Herbert, R.A., van Gemerden, H., Viaroli, P., Cecherelli, V.-U., 447 Rodrı́guez-Valera, F., Bartoli, M., Giordani, G., Azzoni, R., Schaub, B., Welsh, D.T., Donnelly, A., 448 Cifuentes, A., Antón, J., Finster, K., Nielsen, L.B., Pedersen, A.-G.U., Neubauer, A.T., Colangelo, M.A., and 449 Heijs, S.K. (2001) ROBUST: The ROle of BUffering capacities in STabilising coastal lagoon ecosystems. 450 Continental Shelf Research 21(18–19), 2021-2041.
451 452 Duarte, C.M. (2002) The future of seagrass meadows. Environmental Conservation 29(02), 192-206.
453 454 Duarte, C.M., Conley, D.J., Carstensen, J., and Sánchez-Camacho, M. (2008) Return to Neverland: Shifting 455 Baselines Affect Eutrophication Restoration Targets. Estuaries and Coasts 32(1), 29-36.
456 457 Ferguson, A.J.P., Gruber, R.K., Orr, M., and Scanes, P. (2016) Morphological plasticity in Zostera muelleri 458 across light, sediment, and nutrient gradients in Australian temperate coastal lakes. Marine Ecology Progress 459 Series 556, 91-104.
460
18
461 Fiebig, S. (2010) Summary of Ecological Information for the Wallis Lake Potential RAMSAR site. 462 Departmnent of Environment and Climate Change (DECC). New South Wales, Australia, p. 120.
463 464 Fossing, H., and Jørgensen, B.B. (1989) Measurement of bacterial sulfate reduction in sediments: Evaluation 465 of a single-step chromium reduction method. Biogeochemistry 8(3), 205-222.
466 467 Fourqurean, J.W., Duarte, C.M., Kennedy, H., Marba, N., Holmer, M., Mateo, M.A., Apostolaki, E.T., 468 Kendrick, G.A., Krause-Jensen, D., McGlathery, K.J., and Serrano, O. (2012) Seagrass ecosystems as a 469 globally significant carbon stock. Nature Geoscience 5(7), 505-509.
470 471 Frederiksen, M.S., Holmer, M., Borum, J., and Kennedy, H. (2006) Temporal and spatial variation of sulfide 472 invasion in eelgrass (Zostera marina) as reflected by its sulfur isotopic composition. Limnology and 473 Oceanography 51(5), 2308-2318.
474 475 Glud, R.N. (2008) Oxygen dynamics of marine sediments. Marine Biology Research 4(4), 243-289.
476 477 Hansen, J.W., Udy, J.W., Perry, C.J., Dennison, W.C., and Lomstein, B.A. (2000) Effect of the seagrass 478 Zostera capricorni on sediment microbial processes. Marine Ecology Progress Series 199, 83-96.
479 480 Harris, G. (1996) 'Port Phillip Bay environmental study: final report.' (CSIRO Canberra)
481 482 Hasler-Sheetal, H., and Holmer, M. (2015) Sulfide intrusion and detoxification in the seagrass Zostera 483 marina. PLoS One 10(6), e0129136.
484 485 Herbeck, L.S., Sollich, M., Unger, D., Holmer, M., and Jennerjahn, T.C. (2014) Impact of pond aquaculture 486 effluents on seagrass performance in NE Hainan, tropical China. Marine Pollution Bulletin 85(1), 190-203.
487 488 Holmer, M. (2009) Productivity and biogeochemical cycling in seagrass ecosystems. In Coastal Wetlands: 489 An Integrated Ecosystem Approach. (Eds. GME Perillo, E Wolanski, DR Cahoon and MM Brinson) pp. 490 377–401. (Amsterdam, The Netherlands)
491 Holmer, M., Duarte, C.M., Boschker, H.T.S., and Barrón, C. (2004). Carbon cycling and bacterial carbon 492 sources in pristine and impacted Mediterranean seagrass sediments. Aquatic Microbial Ecology 36, 227-237.
493 494 Holmer, M., Georgiev, V.G., and Karamfilov, V.K. (2016) Effects of point source of untreated sewage 495 waters on seagrass (Zostera marina and Z. noltii) beds in the South-Western Black Sea. Aquatic Botany 133, 496 1-9.
497 498 Holmer, M., and Hasler-Sheetal, H. (2014) Sulfide intrusion in seagrasses assessed by stable sulfur isotopes 499 – A synthesis of current results. Frontiers in Marine Science 1:64.
500 501 Holmer, M., and Kendrick, G. (2013) High sulfide intrusion in five temperate seagrasses growing under 502 contrasting sediment conditions. Estuaries and Coasts 36(1), 116-126.
19
503 504 Jensen, S.I., Kühl, M., Glud, R.N., Jørgensen, L.B., and Priemé, A. (2005) Oxic microzones and radial 505 oxygen loss from roots of Zostera marina. Marine Ecology Progress Series 293, 49-58.
506 507 Kennedy, H., Beggins, J., Duarte, C.M., Fourqurean, J.W., Holmer, M., Marbà, N., and Middelburg, J.J. 508 (2010) Seagrass sediments as a global carbon sink: Isotopic constraints. Global Biogeochemical Cycles 509 24(4).
510 511 Kerr, E.A., and Strother, S. (1990) Seasonal changes in standing crop of Zostera muelleri in south-eastern 512 Australia. Aquatic Botany 38(4), 369-376.
513 514 Kohlmeier, D., Pilditch, C.A., Bornman, J.F., and Bischof, K. (2014) Site specific differences in 515 morphometry and photophysiology in intertidal Zostera muelleri meadows. Aquatic Botany 116, 104-109.
516 517 Koren, K., Brodersen, K.E., Jakobsen, S.L., and Kühl, M. (2015) Optical Sensor Nanoparticles in Artificial 518 Sediments–A New Tool To Visualize O2 Dynamics around the Rhizome and Roots of Seagrasses. 519 Environmental Science & Technology 49(4), 2286-2292.
520 521 Marbà, N., Calleja, M., Duarte, C., Álvarez, E., Díaz-Almela, E., and Holmer, M. (2007) Iron additions 522 reduce sulfide intrusion and reverse seagrass (Posidonia oceanica) decline in carbonate sediments. 523 Ecosystems 10(5), 745-756.
524 525 Mascaro, O., Valdemarsen, T., Holmer, M., Perez, M., and Romero, J. (2009) Experimental manipulation of 526 sediment organic content and water column aeration reduces Zostera marina (eelgrass) growth and survival. 527 Journal of Experimental Marine Biology and Ecology 373(1), 26-34.
528 529 Matheson, F.E., and Schwarz, A.M. (2007) Growth responses of Zostera capricorni to estuarine sediment 530 conditions. Aquatic Botany 87(4), 299-306.
531 532 Mellors, J., Waycott, M., and Marsh, H. (2005) Variation in biogeochemical parameters across intertidal 533 seagrass meadows in the central Great Barrier Reef region. Marine Pollution Bulletin 51(1–4), 335-342.
534 535 Møhlenberg, F. (1999) Effect of meteorology and nutrient load on oxygen depletion in a Danish micro-tidal 536 estuary. Aquatic Ecology 33(1), 55-64.
537 538 Oakes, J.M., and Connolly, R.M. (2004) Causes of sulfur isotope variability in the seagrass, Zostera 539 capricorni. Journal of Experimental Marine Biology and Ecology 302(2), 153-164.
540 541 Pagès, A., Welsh, D.T., Robertson, D., Panther, J.G., Schäfer, J., Tomlinson, R.B., and Teasdale, P.R. (2012) 542 Diurnal shifts in co-distributions of sulfide and iron(II) and profiles of phosphate and ammonium in the 543 rhizosphere of Zostera capricorni. Estuarine, Coastal and Shelf Science 115, 282-290.
544
20
545 Peralta, G., Pérez-Lloréns, J., Hernández, I., Brun, F., Vergara, J., Bartual, A., Gálvez, J., and García, C. 546 (2000) Morphological and physiological differences between two morphotypes of Zostera noltii Hornem. 547 from the south-western Iberian Peninsula. Helgoland Marine Research 54(2-3), 80-86.
548 549 Pérez, M., García, T., Invers, O., and Ruiz, J.M. (2008) Physiological responses of the seagrass Posidonia 550 oceanica as indicators of fish farm impact. Marine Pollution Bulletin 56(5), 869-879.
551 552 Robertson, D., Teasdale, P.R., and Welsh, D.T. (2008) A novel gel-based technique for the high resolution, 553 two-dimensional determination of iron (II) and sulfide in sediment. Limnology and Oceanography: Methods 554 6(10), 502-512.
555 556 Roca, G., Alcoverro, T., de Torres, M., Manzanera, M., Martínez-Crego, B., Bennett, S., Farina, S., Pérez, 557 M., and Romero, J. (2015) Detecting water quality improvement along the Catalan coast (Spain) using stress- 558 specific biochemical seagrass indicators. Ecological Indicators 54, 161-170.
559 560 Romero, J., Martinez-Crego, B., Alcoverro, T., and Perez, M. (2007) A multivariate index based on the 561 seagrass Posidonia oceanica (POMI) to assess ecological status of coastal waters under the water framework 562 directive (WFD). Marine Pollution Bulletin 55(1-6), 196-204.
563 564 Roy, P.S., Williams, R.J., Jones, A.R., Yassini, I., Gibbs, P.J., Coates, B., West, R.J., Scanes, P.R., Hudson, 565 J.P., and Nichol, S. (2001) Structure and Function of South-east Australian Estuaries. Estuarine, Coastal and 566 Shelf Science 53(3), 351-384.
567 568 Smith, C.S., and Heggie, D.T. (2003) Benthic nutrient fluxes in Smiths Lake, NSW. Geoscience Australia 569 Record 2003/16.
570 571 Sørensen, J. (1982) Reduction of Ferric Iron in Anaerobic, Marine Sediment and Interaction with Reduction 572 of Nitrate and Sulfate. Applied and Environmental Microbiology 43(2), 319-324.
573 574 Stookey, L.L. (1970) Ferrozine---a new spectrophotometric reagent for iron. Analytical chemistry 42(7), 779- 575 781.
576 577 Udy, J.W., and Dennison, W.C. (1997) Growth and physiological responses of three seagrass species to 578 elevated sediment nutrients in Moreton Bay, Australia. Journal of Experimental Marine Biology and Ecology 579 217(2), 253-277.
580 581 Waycott, M., Duarte, C.M., Carruthers, T.J., Orth, R.J., Dennison, W.C., Olyarnik, S., Calladine, A., 582 Fourqurean, J.W., Heck, K.L., Jr., Hughes, A.R., Kendrick, G.A., Kenworthy, W.J., Short, F.T., and 583 Williams, S.L. (2009) Accelerating loss of seagrasses across the globe threatens coastal ecosystems. 584 Proceedings of the National Academy of Sciences of the United States of America 106(30), 12377-81.
585
586 587
21
588 Figures legends 589
590 Fig. 1 Map of sampling stations in Wallis Lake. Modelled water column concentrations of TN (µg l-1) are 591 shown for the study area (Fiebig 2010)
592 Fig. 2 Sediment characteristics in Wallis Lake. (A) sediment organic carbon content (POC, %DW), (B) 593 sediment organic nitrogen content (PON, %DW), sediment C:N ratio (molar) and sediment sulfides (TRS ; 594 µmol cm-3). Letters indicate significant differences (ANOVA and Tukey´s test p>0.05; n=5, ±SEM).
595 Fig. 3 Two DGT samplers from Sta. 1 and Sta. 4 showing precipitation of porewater sulfides in greyscale. 596 Areas with no sulfide precipitation (white) are found around roots of Z. muelleri.
597 Fig. 4 Seagrass shoot density (A, shoots m-2) and above- and below-ground biomass (B, g DW m-2) from 598 lower estuary (Sta. 1) to lagoon (Sta. 4 and 5) in Wallis Lake. Letters indicate significant differences 599 (ANOVA and Tukey´s test p>0.05; n=5, ±SEM).
600 Fig. 5 Relationship between total sulfur content (TS, %DW) in the plant tissue and Fsulfide (%) in Z. muelleri 601 in Wallis Lake. Best linear fit and correlation coefficients are given.
602
22
32°05’
Forster 32°10’
Pipers Creek 32°15’ Lake Wallis Marine lagoon 32°20’ New South Wales
152°25’ 152°30’ 152°35’ 8 0.8 A c c B c 6 0.6 bc b b 4 0.4 POC (%) POC PON (%) PON 2 a 0.2 a a a 0 0.0 1 2 3 4 5 1 2 3 4 5
80 150 d C b D
60 c ) 3 100 c 40 b
C:N (%) C:N a 50 a a 20 a a TRS ( µ mol cm
0 0 1 2 3 4 5 1 2 3 4 5
site i n c r e a s i n g s e d i m e n t d e p t h 1 c m A B 80 600 A b B b ab
400 60 )
-2 a a )
-2 200 a 40 a a 0 a
Biomassm DW (g ab Shootdensity (shootsm 20 a a -200 ab ab
b 0 -400 1 2 3 4 5 1 2 3 4 5 Site Site 100 Leaves R2=0.58 Rhizomes 80 Roots
2 (%) 60 R =0.72
sulfide 2 F 40 R =0.51
20
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 TS (%) 2- 34 Table 1: Sediment Fe content, porewater sulfate (SO4 ) from DGT gels and δ S signals in porewater sulfate and sediment AVS and CRS in Lake Wallis (mean±SD; n=4-5). The contribute of seagrass detritus to sediment POC pool (%seagrass) is calculated as average by use of IsoSource (see text for explanations).
-3 2- 34 34 34 Fe (mmol cm ) SO4 (mM) δ S Sulfate (‰) δ S AVS (‰) δ S CRS (‰) %seagrass Sta. 1 67.26±4.71 21.2±5.2 20.84±0.43 -22.53±1.65 -28.09±0.20 7.5 Sta. 2 58.97±9.76 21.9±9.7 21.09±0.20 -22.30±2.37 -27.81±0.57 28 Sta. 3 73.95±7.59 31.6±1.1 20.99±0.12 -20.01±1.03 -27.09±0.26 29 Sta. 4 59.56±14.86 19.9±3.4 21.19±0.28 -23.33±1.16 -25.02±0.54 75 Sta. 5 69.77±14.30 19.2±4.2 21.11±0.14 -19.82±1.96 -23.49±1.19 71
Table 2: Seagrass morphological parameters in Lake Wallis Letters indicate significant differences (ANOVA and Tukey’s test p < 0.05); (n = 5, ±SD)
A:B ratio Shoot weight Leaf length Leaf width Flowering shoots (mg DW-1) (mm) (mm) Sta. 1 1.80±0.47 134±51 b 365±76 b 3.0±0.8 ac 48±50 ab Sta. 2 1.31±0.61 114±15 b 299±67 c 3.2±0.4 bc 36±33 b Sta. 3 1.40±0.62 113±50 b 305±88 c 3.0±0.4 c 36±80 b Sta. 4 1.34±0.45 153±75 b 358±71 b 3.5±0.4 a 167±115 a Sta. 5 1.86±0.40 276±78 a 580±130a 3.4±0.4 ab 0 b
13 15 Table 3: Average carbon (%C), nitrogen (%N), stable isotopic signal (δ C, δ N ‰), Fsulfide, C:N ratio (molar) and sulfur (TS) content (%DW) sediment of Z. muelleri leaves, rhizomes and roots in Lake Wallis, respectively total reducible sulfide *(TRS) in Lake Wallis sediments (n=5, mean±SD). Lower case letters indicate significant differences between stations. ANOVA and post hoc Tukey’s test; p < 0.05.
%C (%DW) δ13C (‰) %N (%DW) δ15N (‰) C:N (molar) Fsulfide TS (%DW)/TRS (µmol cm-3) Leaves Sta. 1 34.32±0.83ns -11.43±1.12a 1.54±0.09b 2.71±0.49b 26.0±1.7a 46.4±1.9c 0.62±0.06b Sta. 2 34.60±1.26ns -9.50±0.44b 1.54±0.07b 2.95±0.38b 26.3±1.7a 39.4±2.9bc 0.59±0.10ab Sta. 3 33.82±0.80ns -10.32±0.44ab 1.41±0.13ab 4.88±0.16c 28.2±3.2ab 48.3±4.6c 0.76±0.11c Sta. 4 33.82±0.56ns -9.61±0.57b 1.30±0.13a 1.14±0.83a 30.6±2.8b 18.0±1.7a 0.52±0.10ab Sta. 5 33.11±0.60ns -9.88±0.44b 1.29±0.12a 0.95±0.61a 30.2±2.8b 31.5±2.2b 0.49±0.04a Rhizomes Sta. 1 36.48±1.52ns -11.48±0.30a 0.73±0.23b 2.38±1.23ns 63.9±21.7ns 55.0±5.0ab 0.66±0.22ns Sta. 2 34.72±2.02ns -10.32±0.84b 0.45±0.12a 4.16±3.05ns 95.7±26.0ns 57.4±4.3b 0.61±0.16ns Sta. 3 37.14±0.69ns -10.68±0.28ab 0.53±0.05ab 4.35±0.33ns 82.4±7.9ns 39.9±2.6a 0.33±0.06ns Sta. 4 36.69±0.94ns -9.85±0.77b 0.39±0.11a 2.33±4.26ns 118.1±39.9ns 55.2±7.4ab 0.54±0.26ns Sta. 5 36.57±1.01ns -10.11±0.41b 0.75±0.16b 1.54±0.16 ns 59.7±15.3ns 49.3±4.9ab 0.46±0.11ns Roots Sta. 1 34.92±1.37ns -11.04±0.25ac 0.85±0.11a 2.12±0.63ns 48.7±7.3ns 72.1±2.0ab 0.77±0.13ab Sta. 2 34.60±2.04ns -9.70±0.57c 0.70±0.09a 2.99±0.89ns 58.0±6.8ns 57.5±3.2a 0.62±0.12a Sta. 3 32.37±2.17ns -11.41±0.30a 0.66±0.12a 4.28±0.95ns 58.5±10.4ns 80.5±3.0b 1.03±0.19b Sta. 4 35.00±0.75ns -10.37±0.53abc 0.79±0.03a 1.49±0.42ns 51.7±2.3ns 71.7±3.9ab 0.83±0.24ab Sta. 5 31.78±3.63ns -10.21±1.07bc 0.76±0.13a 2.54±2.45ns 50.6±15.8ns 79.0±5.3b 0.72±0.26ab Sediment Sta. 1 3.48±0.21b -18.05±0.23c 0.35±0.03b 1.64±0.03a 11.6±1.2a nd *133.5±5.7d Sta. 2 0.97±0.16a -16.04±0.08b 0.05±0.02a 1.05±0.48a 21.9±1.7c nd *51.2±2.2b Sta. 3 0.76±0.14a -16.34±0.3b 0.02±0.01a 3±0.36b 58.0±17.1d nd *28.2±4.7a Sta. 4 6.82±0.23c -14.63±0.15a 0.57±0.03c 1.25±0.04a 13.6±0.1b nd *78.8±6.2c Sta. 5 6.71±0.15c -14.02±0.21a 0.51±0.04c 1.39±0.04a 15.2±0.4b nd *94.8±4.5c
Table 4: Average C (%C) and N (%N) content (%DW), stable isotopic signal (δ13C, δ15N ‰) and C:N ratio (molar) of potential carbon sources in sediments of Lake Wallis. n= number of replicates.
n %C (%DW) δ13C %N (%) δ15N C:N Origin Halophila sp. 2 30.46 -14.03 1.90 0.09 18.6 Seston 2 11.55 -18.10 0.90 4.39 14.9 Macroalgae 1 23.73 -18.10 1.32 2.74 21.0 Seagrass detritus 5 32.60 -11.79 0.78 2.32 51.1
Table 5: Linear regression of Fsulfide in leaves versus seagrass morphology and sediment biogeochemistry
Equation R2 p Shoot density Y = -0.005510*X + 45.94 0.21 0.0243 Above-ground biomass Y = -0.04267*X + 47.51 0.38 0.0014 Below-ground biomass Y = -0.04252*X + 43.90 0.20 0.0273 A:B Y = -1.964*X + 39.29 0.006 0.7091 Leaf width Y = -34.46*X + 147.9 0.48 0.0002 Leaf length Y = -0.02984*X + 47.62 0.23 0.2272 %N leaves Y = 49.14*X - 32.84 0.32 0.0037 δ15N Y = 5.693*X + 21.78 0.49 0.0001 %C seagrass Y = -0.3658*X + 51.78 0.62 <0.0001 sedPOC Y = -3.323*X + 48.98 0.48 0.0002 sedPON Y = -34.91*X + 47.22 0.41 0.0007 AVS Y = -6.718*X + 46.09 0.40 0.0009 CRS Y = -0.006269*X + 36.69 <0.001 0.9309