Effects of suspended sediments on the sponge holobiont with implications for dredging management

Dredging can cause high suspended sediment concentrations (SSC) in the water column, posing a hazard to filter feeding organisms like sponges as sediment may clog their aquiferous systems and reduce feeding. In order to provide pressure−response values for sponges to SSC and tease apart the cause:effect pathways of dredging pressures, five heterotrophic and phototrophic species were experimentally exposed to a range of dredging-relevant SSC of up to 100 mg L−1, with light compensation across treatments to ensure that SSC was the primary physical parameter. This study shows that some sponge species exposed to high SSC (≥23 mg L−1) for extended periods (28 d) have lower survival, increased necrosis and depletion of energy reserves. In contrast, SSC of ≤10 mg L−1 caused few, if any, negative effects and is thus suggested as a prudent sub-lethal threshold for sponges. Microbial communities did not change significantly among SSC treatments, although a nutritional shift from mixotrophy towards increased phototrophy was detected for some sponge species exposed to high SSC. Importantly however, it is expected that the combined effect of SSC with low light availability and sediment smothering as occurs during dredging operations will increase the negative effects on sponges.

especially in bioeroding sponge species 19,25,26 . In some tropical sponge species the photosymbionts can provide >50% of the energy requirements of the host 18,27 , and the diverse symbiotic consortia can comprise up to 40% of the sponge volume 28 . Sponge symbionts tend to be highly host specific, but are generally stable across broad geographic and environmental gradients 29 , with the stable host-microbe consortium often referred to as the 'sponge holobiont' 28 .
Depending on their degree of nutritional dependence from symbiont primary production, the sponge holobiont can be described as either 'phototrophic' , 'mixotrophic' or 'heterotrophic' . Some species exhibit considerable flexibility in their feeding strategy and are able to alter their nutritional mode depending on prevailing conditions 27,30,31 . The different modes of nutrition have important implications for understanding the environmental effects of dredging activities on sponges. Due to their filter feeding activity, sponges may be affected by elevated suspended sediment concentrations (SSCs) and long term exposure to high SSCs can lead to clogging of their aquiferous systems, reducing the flow of oxygenated seawater to the sponge mesohyl and compromising heterotrophic feeding 32 . Some sponges exhibit short term responses to high turbidity such as temporarily closing or reducing the size of their incurrent openings (ostia) and arresting pumping (filtering) activity [33][34][35][36] . However, reduced pumping activity has flow-on consequences for host energetics, health and reproductive output 32,[37][38][39] . Since some sponges also rely on photoautrotrophy for nutrition, they may be affected by the pronounced reduction in benthic light availability that occurs in dredging plumes 4,40,41 . Periods of low light levels associated with dredging can cause photoacclimation of the photosymbionts 32,42,43 , but extended periods of low light can also result in dissociation of the holobiont (bleaching) and subsequent mortality in some phototrophic species 44 .
Understanding the effects of turbidity on sponges would facilitate development of water quality thresholds (or trigger values) that could be used to alert dredging proponents of conditions that could, if continued, cause environmental harm. Depending upon permit requirements, the option could then be to modify the dredging operations i.e. changing the rate or location of dredging, and hence avoid or minimize impacts. Developing thresholds is predicated upon identifying key cause-effect pathways and determination of dose−response relationships which is technically challenging. Close to dredging activities, sponges will be exposed to both high SSCs in combination with low light conditions (caused by the water turbidity), and may also be simultaneously exposed to high levels of sedimentation and possibly smothering. There are, therefore, many potential cause-effect pathways. In order to avoid conflation of these in a single experiment (see ref. 4), in this study we examine the response of sponges to a range of suspended sediments but kept light uniform across treatments. Experiments were conducted over 28 d, with a similar suite of 3 heterotrophic and 2 phototrophic sponges as used in previous studies assessing the effects of sedimentation 38 and light 44 .
Sponge growth, necrosis and mortality. Most sponges appeared visibly healthy at the end of the experiment, with the exception of all Carteriospongia foliascens, a few individuals of Coscinoderma matthewsi, and Cliona orientalis exposed to 100 mg L −1 . No visible effects were observed in Cymbastela coralliophila and Stylissa flabelliformis in any treatment. Percentage growth and relative growth rates based on thickness measurements before and after the 28 d sediment exposure and observational period, were negative for most sponges exposed to SSC of ≥30 mg L −1 ; however with the exception of Cliona orientalis, these negative growth rates were not significantly different from control groups (Fig. 2a, Table 1a). The bioeroding sponge Cliona orientalis grew during the sediment exposure and recovery periods in all treatments except at the highest SSC of 100 mg L −1 (Fig. 2a, Table 1a). The massive sponge Coscinoderma matthewsi shrank when exposed to suspended sediments (3-100 mg L −1 ) and only grew in the control treatment, although this growth was not statistically significant. The erect sponge Stylissa flabelliformis shrank in all treatments with tissue regression increasing with increasing SSC and no growth evident during the observational period (Fig. 2a, Table 1a). Sponge growth based on surface area  Table S1).
The percentage of dead (i.e. partial mortality or necrosis) and bleached (discoloured) tissue also increased significantly with higher SSC in all species except for Cymbastela coralliophila and Stylissa flabelliformis (Fig. 2b, Table 1b). The highest percentages of necrosed tissue were observed among individuals of Carteriospongia foliascens and Coscinoderma matthewsi exposed to the highest SSC, reaching 80 and 60% of sponge tissue respectively by the end of the observational period. Carteriospongia foliascens showed signs of necrosis from the second week of sediment exposure to 100 mg L −1 , although some necrosis was also observed in lower treatments by the end of the experiment (Fig. 2b). In terms of total mortality, 90% of Carteriospongia foliascens died in the 100 mg L −1 treatment by the end of the observational period and 20% had died in each of the 30 and 10 mg L −1 treatments ( Supplementary Fig. S2). The bioeroding sponge Cliona orientalis also showed signs of stress, as evidenced by significant bleaching, within the 100 mg L −1 treatment, although significant recovery of pigmentation occurred during the observational phase (ANOVA: P < 0.05). Coscinoderma matthewsi also exhibited considerable necrosis at high SSC including 20% total mortality in the 100 mg L −1 treatment by the end of the observational period ( Supplementary Fig. S2). No mortality occurred in Cymbastela coralliophila, Cliona orientalis and Stylissa flabelliformis.
Mortality data for Carteriospongia foliascens was fitted to nonlinear regression curves to calculate the lethal concentration (LC) at which 50% (LC 50 ) and 10% (LC 10 ) of the population died. Nonlinear regression of the dose−response curve (R 2 = 0.8675, AICc = 85.23) met assumptions of normality and homoscedasticity and there was no evidence for a lack of fit in the replicates test (P = 0.694). After the 28 d exposure period, the LC 50 (and 95% confidence intervals range) for mortality in Carteriospongia foliascens was 40.6 mg L −1 (range: 28.9-57.0 mg L −1 ) and the LC 10 was 21.5 mg L −1 (range: 13.1-35.2 mg L −1 ). Mortality data from Coscinoderma matthewsi did not meet model assumptions and LC values could not be calculated. Asterisks show statistically significant differences between the experimental and observational phase in a-c and between treatments in d (t-tests: P < 0.05).
Scientific RepoRts | 7: 4925 | DOI:10.1038/s41598-017-05241-z Lipids. For all species, total lipid content was similar between time of sample collection and the start of the sediment dosing, as well as during the experiment for sponges exposed to no suspended sediment (ANOVA: P > 0.05). With the exception of Cymbastela coralliophila, all species had lower lipid content when exposed to SSCs ≥ 30 mg L −1 , although this was only statistically significant for Cliona orientalis at the end of the 28 d sediment exposure (Fig. 2c, Table 1c). Unexpectedly, Cliona orientalis had a higher lipid content in the 100 mg L −1 treatment than in the 30 mg L −1 treatment and while lipid content appeared to decrease in samples from the 100 mg L −1 treatment during the observational phase, this decrease was not significant. Coscinoderma matthewsi and Stylissa flabelliformis increased their lipid content during the observational period (Fig. 2c).
Respiration rates. Respiration rates did not significantly differ among treatments at day = 0 for any species (Fig. 2d, Table 1d). However, after 28 days, respiration rates were significantly lower for Carteriospongia   foliascens and Cliona orientalis exposed to suspended sediment than conspecifics in the 0 mg L −1 treatment (ANOVA, P < 0.05; Fig. 2d). In Carteriospongia foliascens, respiration rates were lower than controls in the 10 and 30 mg L −1 treatments and the absence of a significant difference at 100 mg L −1 is likely due to the high mortality of Carteriospongia foliascens in this treatment (which precluded post hoc analysis, Table 1d). No interactive effect of time and treatment was observed for Carteriospongia foliascens. Respiration in Cliona orientalis in the 0 mg L −1 treatment increased significantly between day 0 and the end of the experimental phase, returning to pre-treatment levels in the observational period. Although this response in control sponges was unexpected, respiration rates were still observed to decrease for Cliona orientalis in all sediment treatments over time. Furthermore, respiration rates were significantly lower at the end of the observational phase than at day 0 within each sediment treatment, indicating that a longer recovery period was required for Cliona orientalis to return to pre-exposure respiration rates (Fig. 2d, Table 1d). A similar recovery capacity was observed in Cymbastela coralliophila, although increased respiration during the observational phase was only detected in the 3 mg L −1 treatment. Respiration rates in the massive species Coscinoderma matthewsi were significantly increased in sponges in the 100 mg L −1 treatment.
Coscinoderma matthewsi was the only species to exhibit an increased respiration rate on exposure to SSC. Finally, respiration rates in Stylissa flabelliformis were not significantly different among treatments, although when looking within each sampling time individually, a significant decrease at the end of the 28 d was detected for all treatments exposed to sediments, with recovery only evident in the sponges exposed to the 3 mg L −1 treatment (Fig. 2d, Table 2d).
Histology. Choanocyte chambers could only be accurately quantified in Cliona orientalis. While the number of choanocyte chambers at the end of the experimental period was 37.3 ± 11.6 and 9.7 ± 2.7, (mean ± SD) in sponges from the 0 mg L −1 and 100 mg L −1 treatments respectively, this difference was not statistically significant (t test: P = 0.103)( Supplementary Fig. S3).

Chlorophyll fluorescence.
Overall, no significant differences in maximum quantum yield were observed between treatments for any of the phototrophic species hosting cyanobacterial symbionts, with minor fluctuations throughout the experiment and observational periods (Fig. 3a,   but while the decrease with time was significant, the interaction of time and treatment was not (Fig. 3a). In contrast, for Cliona orientalis which hosts Symbiodinium symbionts, the interaction term (treatment*time) was significant. Within the 10 mg L −1 treatment, quantum yields decreased significantly on day 14 and were also lower on observation day 7 than on day 7 of the exposure (Fig. 3a, Table 2a). Within the 100 mg L −1 treatment, quantum yields decreased after 7 days and then significantly increased after 21 days. After 21 days exposure, quantum yields in sponges from 0 and 10 mg L −1 treatments were significantly different from sponges in the 100 mg L −1 treatment. After 14 days of the observational period, quantum yields were significantly greater than in sponges exposed to 100 mg L −1 for 21 days.
Pigment analysis. Chl a, was highly correlated with Total chlorophyll for Carteriospongia foliascens, Cymbastela coralliophila and Cliona orientalis (R 2 > 0.85, P < 0.001), so Chl a concentration was used as a proxy for total photosymbiont health (i.e. bleaching). Overall, concentrations of Chl a were stable throughout the experiment (Fig. 3b). Significant differences in Chl a among SSC treatments were only observed for Carteriospongia foliascens at the end of the 28 d sediment exposure period, with lower concentrations in individuals exposed to 100 mg L −1 (Table 2b). However, a negative and significant correlation between Chl a and Chl d in both Carteriospongia foliascens and Cymbastela coralliophila (R 2 = −0.254, −0.265, and P = 0.046, 0.037, respectively), and an increase in Chl d in samples exposed to 100 mg L −1 in both species ( Supplementary Fig. S4), suggests an increase in some Chl d-containing Cyanobacteria under high SSCs. Chl a concentration in samples exposed to high SSCs returned to control levels following the observational period for all three species (Fig. 3b). Non-metric Multi-Dimensional Scaling (nMDS) analysis of normalized data for all pigments retrieved by spectrophotometry (Chl a, Chl b, Chl c, Chl d, Total Chlorophyll and Carotenoids) showed no grouping according to SSC treatment (Fig. 3c). The only exception was Carteriospongia foliascens exposed to 100 mg L −1 , which grouped closer together, consistent with patterns observed for Chl a. PERMANOVA analysis confirmed significant differences between treatments in Carteriospongia foliascens, but also in Cliona orientalis, with subsequent pair-wise testing showing main differences between low and high SSCs (Table 2c).

Microbial community analysis.
A total of 9,530,650 high quality 16 S rRNA gene amplicon sequences were recovered from the 5 sponge species (n = 200 individual samples) and 9 seawater samples. Each sponge species maintained a unique microbial community (Table 3a, Fig. 4, see also Supplementary Fig. S5 and Table S2) that was distinct from the seawater microbiome (Table 3b, Supplementary Figs S5 and S6). Aquarium acclimation (for 4 weeks) did not affect the sponge-associated microbial community of any species (Table 3a) as field controls were not significantly different from Time 0 controls (i.e. samples collected after the acclimation period, right before starting the exposure to sediments) (Table 3a). However, both field/time 0 controls showed significant differences with samples at the end of the experimental/observational period in all species (Table 3a). Elevated SSC did not have a major impact on the overall composition of the sponge microbiome at the phyla level (Fig. 4b) and no significant differences were detected at the OTU level (97% sequence similarity) for any species (Table 3c). In addition, no clear groupings were observed in the ordination (Fig. 4a). The only exception to this microbial stability was in Carteriospongia foliascens, where samples exposed to 100 mg L −1 exhibited greater dispersion across the ordination (Fig. 4a). Cliona orientalis also showed greater separation between samples at the end of the experimental and observational periods, particularly in the 30 and 100 mg L −1 treatments (Table 3c, Fig. 4a).
Network analyses of the 30 most discriminatory operational taxonomic units (OTUs) for the control (0 mg L −1 ) and highest SSC (100 mg L −1 ) in each species revealed phylogenetically diverse and stable core microbiomes (i.e. the component shared between all samples) as well as treatment specific OTUs (Fig. 4c, Supplementary  Table S2). In each species, the number and taxonomy of OTUs observed exclusively in control samples was generally balanced by the number and taxonomy of new OTUs that were observed only in sediment treated samples. A notable exception however was the recruitment of 3 novel Cyanobacteria OTUs in Carteriospongia foliascens within samples from the high SSC. Also of interest was the finding that core Cyanobacteria OTUs in the two photosynthetic species Carteriospongia foliascens and Cymbastela coralliophila are not disrupted by sediments, even at the highest SSC.

Discussion
Establishing dose-response relationships relies on identifying the significance and relative importance of different cause-effect pathways. As part of a sequence of experiments, and a wider investigation into the effects of sediment deposition, light attenuation and elevated SSC (alone and in combination) 38, 44 , we exposed five sponge species spanning a range of morphologies and nutritional modes to elevated SSCs for a chronic (28 d) period. Light levels were kept constant across the different treatments to isolate SSC as the primary variable. The SSCs used in this study (range 0-100 mg L −1 ) were designed to bracket SSCs measured during a recent large scale capital dredging project in NW Australia (the Barrow Island project, see ref. 4). During this project, at 4 shallow water sites (~5-11 m depth) located <~0.5 km from the dredging, the 95 th percentile (P 95 ) of SSCs over a 30 d running mean period was 16.0 mg L −1 (range: 12.9-20.9) during the dredging phase as compared to 2.0 mg L −1 (range: 1.5-3.1) before dredging. Exposure of the sponges to SSCs ≥23 mg L −1 which corresponded to the 30 and 100 mg L −1 nominal treatments, for 28 d had an overall negative effect on sponge physiology, including decreases in lipid content, suggesting that sponge feeding behaviour may be compromised. High levels of SSC also resulted in negative growth, decreased respiration rates and caused significant necrosis and mortality within the cup sponge Carteriospongia foliascens and, to a lesser extent, the massive sponge Coscinoderma matthewsi. The high levels of Carteriospongia foliascens mortality in response to elevated SSCs supports previous research proposing this species as a sensitive bioindicator for assessing impacts from the dredging related pressure of light attenuation 44 .
Cliona orientalis was also negatively affected by high SSC, although it demonstrated a capacity for rapid recovery. LC 50 and LC 10 levels could only be calculated for Carteriospongia foliascens and were 40.61 and 21.51 mg L −1 , respectively. On the other hand, exposure to SSCs ≤ 10 mg L −1 was tolerated by most species over the experimental period. These water quality values cannot be used for impact prediction and management purposes in isolation, as the SSCs are likely to have profound effects on light attenuation (see refs 4, 40 and 41) which is also known to significantly affect some sponge species used in this study 44 .
In general, respiration rates decreased with elevated SSC, consistent with responses reported for the deep sea sponge Geodia barretti following sediment exposure 45,46 . Cliona orientalis, Cymbastela coralliophila and Stylissa flabelliformis were able to metabolically recover within two weeks when exposed to the lowest sediment treatment (3 mg L −1 ); however, a greater recovery time is likely required at higher SSC. In contrast to the other species, respiration rates in the massive species Coscinoderma matthewsi increased significantly at 100 mg L −1 , a trend that has also been reported in other heterotrophic massive species following short term exposure to elevated SSC 47 . However, the higher respiration rates in Coscinoderma matthewsi may also be a result of stress associated with oxygen depletion within the experimental respiration chambers 48 . Hence, due to the wide range of species-specific responses, it is difficult to generalise about how sponge respiration rates will respond to elevated SSCs.
Sponges are highly efficient filter-feeders, processing large amounts of water from which they efficiently remove bacteria and nutrients 13,14 . Elevated SSCs from natural events or anthropogenic activities such as dredging are a natural hazard to the filter feeding mode of nutrition 49,50 . Different sponge species have evolved different strategies to avoid excessive sediment intake and prevent clogging of their aquiferous systems. Sponges can slow their pumping activity 33,34,46 and accumulate sediments in pockets that can be subsequently expelled with mucus secretions (reviewed in refs 32 and 39). Some sponges such as Cliona orientalis, are also capable of entirely closing their oscula (excurrent opening) in response to sediment deposition 36 . Oscular closure can decrease and entirely arrest sponge pumping within hours 36 which would result in decreased respiration rates over very short time scales 46,51 .
Different sediment removal mechanisms are likely to have very different energetic demands. However, all strategies are thought to contribute to a reduction in the food-retention efficiency of the sponges' aquiferous system 47, 52 which may compromise energy stores in the long term. Lipids are essential to an organism's physiological processes as they are a major source of metabolic energy and are often considered the most effective energy store in marine ecosystems 53,54 . Lipids can be broadly grouped into storage and structural functions, with storage lipids predominantly being used as energy reserves and both categories playing a significant role in stress tolerance [55][56][57] . A decrease in lipids has previously been associated with lower levels of energy storage and lower survival rates in corals 58, 59 and sponges 16 . However, in this study, no significant decreases in total lipid reserves were observed in sponges exposed to high SSC. Phototrophic sponges may have the potential to compensate their energy intake through photosymbiosis. For instance, the variable lipid content in the phototrophic Cliona orientalis across treatments likely reflects the flexible feeding strategy of this species including an ability to switch to heterotrophic feeding, as evidenced by its survival during extended periods of darkness 44 and consistent with results for other species of the same genera 60 . The increase in total lipids in Cliona orientalis in the 100 mg L −1 treatment may result from an increase of the photosynthetic dinoflagellates Symbiodinium sp. in an attempt to increase phototrophic feeding, as corroborated by histological results. In a similar way, Symbiodinium sp. living in shallow water corals have been reported to supply host polyps with an excess of energy-rich lipids 61,62 . In Cliona orientalis these energy stores were likely used during the observational period to aid recovery from bleaching and tissue regression and to increase choanocyte chamber density; although a longer recovery time may be necessary to return total lipids to control levels.
As light was compensated across treatments to maintain equivalent daily light integrals and intensities irrespective of SSC, no direct effects were detected of SSC on maximum quantum yields or Chl a concentrations, until disruption of the holobiont due to temporary (e.g. Cliona orientalis) or permanent bleaching (e.g. Carteriospongia foliascens). Hence, our results suggest that the health and photosynthetic efficiency of the photosymbionts is unaffected by high SSC in the absence of light attenuation. Interestingly though, increased chlorophyll d in samples of Carteriospongia foliascens and Cymbastela coralliophila exposed to 100 mg L −1 suggested an increase in some Chl d-containing Cyanobacteria 63 , providing further support for increased reliance on phototrophic feeding under high SSC. Previous research has also shown that the presence of photosymbionts often influences the composition of the host associated microbiome 64 hence any SSC induced change in photosymbiosis may have flow-on effects for the entire sponge microbiome.
Overall, sponge associated microbiomes were not significantly affected by SSC in any species, although minor differences were observed in Carteriospongia foliascens and Cymbastela coralliophila. In particular, under high SSC (100 mg L −1 ) an increase in Firmicutes, Deltaproteobacteria and Epsilonproteobacteria was evident which is consistent with numerous reports of microbial shifts associated with stressed and diseased sponges [65][66][67] . Interestingly, novel Cyanobacteria OTUs inhabited Carteriospongia foliascens and Cliona orientalis exposed to high SSC, indicating a potential mechanism for acclimation to a high SSC environment. However, while some species may have the ability to successfully alter the composition of their microbial community under different environmental conditions 44,68,69 , species with intimate and/or potentially obligate symbioses can be adversely impacted by disruption of their microbiome 66,67 or loss of symbiotic function 70 . The rapid deterioration in health of Carteriospongia foliascens under conditions of high SSC following the microbiome shift is consistent with its documented intimate reliance on a highly specialised microbial community 44 .
Our results and previous studies suggest that most sponges have well developed mechanisms to survive periods of elevated SSC, however extended exposure to high SSC has significant adverse effects on sponge metabolism 34,45,47 . Some species appear to have plasticity in modifying their aquiferous system 71,72 , including the density of their choanocyte chambers 73 depending on environmental conditions. Shifts from phototrophy to heterotrophy depending on irradiance have been also described in numerous sponge species 16,27,74 . Alternatively, mixotrophic sponges may increase their reliance on phototrophic nutrition as observed in the current study. Our results also suggest that discolouration (i.e. bleaching) and necrosis of sponge tissue are effective bioindicators for dredging related stress.
In this study, the isolation of SSC from other dredging-related factors (i.e. sedimentation and light attenuation) allowed us to identify a specific cause:effect pathway and determine lethal and sub-lethal SSC for sponges 4 , including LC 50 and LC 10 levels for Carteriospongia foliascens (i.e. 40.61 and 21.51 mg L −1 , respectively). In conjunction with previous research on the impacts of light attenuation 44 , these thresholds can now be used in water quality monitoring programs to alert dredging proponents to levels of light reduction and SSC that, if continued, could detrimentally impact sponge populations. In combination with sediment plume and light attenuation models, these results can be used to predict the likely effects of dredging (i.e. at the EIA stage) and to establish impact zones. By modelling different dredging scenarios (i.e. volume of material dredged, tidal phase, over flow options, etc.) the information could also be used to identify optimal dredging scenarios that minimise the likelihood and extent of impact.
Here we show that exposure to high SSC (≥23 mg L −1 ) for extended periods (28 d) has a negative effect on sponge feeding behaviour with associated depletion of energy reserves. However, while ≤10 mg L −1 for <28 d seems tolerable by most species and could be established as a prudent sub-lethal threshold in adult sponges, it is expected that the combined effect of SSC with low light availability and sediment smothering will increase the negative effects on sponges under realistic field conditions. These experimental results will assist regulators and environmental managers in reducing risks from dredging development although experimental research combining sedimentation, SSC and light attenuation is required before final thresholds can be derived for dredging impacts on sponges. Sponge explants were exposed to 5 nominal SSCs of 0 (i.e. experimental control), 3, 10, 30 and 100 mg L −1 and the tanks were supplied with a continuous inflow of 5 µm filtered seawater at a rate of 400 mL min −1 , providing ~6 complete turnovers of seawater in each tank per day. Each treatment level had 3 replicate tanks containing 3-4 sponge replicates per species. To replace sediment lost from the tanks because of the continuous flow through of filtered seawater, new sediment was periodically introduced by small volumes of sediment slurry (~8 g L −1 ) from an adjacent 500 L stock suspension. SSCs within each tank were monitored and controlled from turbidity readings (as nephelometric turbidity units, NTUs) using Turbimax CUS31 (Endress and Hauser, Germany) nephelometers connected to a programmable logic controller (PLC) system (see below). NTUs were converted to SSCs (as mg L −1 ) by applying sediment specific algorithms determined gravimetrically (filtration of SSCs through 0.4 µm (nominal pore size) polycarbonate filters and calculating dry weight of the filters). The PLC system monitored NTU in each tank and controlled the SSCs by episodically opening and closing solenoid valves connected to the stock tank. Triplicate 250 mL water samples were collected from each tank weekly and the SSCs determined gravimetrically. Differences in SSC between treatments throughout the experiment were studied at each sampling day separately with a one-way analysis of variance (ANOVA) using treatment as the fixed factor. Data was log-transformed to achieve homogeneity of variances and normality when required.

Sample collection. To assess impacts across nutritional modes, this study used the phototrophic sponges
The sediment used in this experiment was selected after an initial comparison between inshore sediments (siliclastic sediment collected subtidally from Onslow, Western Australia) 38  In order to isolate the effect of SSC from the associated light attenuation that would occur under different SSCs, light intensities were adjusted between the tanks using a Li-250A light meter (LI-COR Biosciences, NE, US) at the grid level and adjusting the light intensity provided by the LED lights accordingly to ensure that all sponges received the same DLI regardless of SSC treatment. The surface of all sponges was also gently brushed each day to remove any deposited sediment that could negatively affect the sponges. The SSC experiment was conducted for a 28 d 'experimental' period, followed by a 14 d 'observational' period in clear water (0 mg L −1 ). The length of the experiment was selected to simulate chronic conditions and to facilitate comparison with the 30-days running means for dredging-related turbidity events from 3 major capital dredging programs in North Western Australia 41 .
Studied parameters and statistical analyses. The effect of SSC on the sponge holobiont was determined using a suite of response variables, with a particular focus on changes in sponge feeding strategies, including changes in sponge photosymbionts, composition of the sponge microbial community and growth. To obtain baseline data on sponge health, 6 extra individuals were processed for each species immediately after collection (field controls) and after aquarium acclimation (t = 0 controls). From the initial 10 individuals stocked per species and treatment, 5 individuals were sampled (and removed from the experiment) for 'destructive' response variables (i.e. pigments, lipids and microbial community analyses) at the end of the experimental period and the remaining 5 individuals were sampled at the end of the observational period, in order to avoid negative effects on sponge health due to sub-sampling of the same individuals. Unless otherwise stated, statistical analyses were performed and graphs prepared using the software R v. 3.1.0 and SigmaPlot v.11.0 (Systat Software Inc.).
Linear mixed models using tank as a random factor and treatment as a fixed factor were fitted by residual maximum likelihood (REML) for data pertaining to relative growth rates, necrosis, lipid content, respiration, histology, chlorophyll fluorescence and total pigments at the end of the exposure and observational periods for each species, separately. An analysis of variance (ANOVA) table was generated for each model and Tukey's post-hoc multiple comparisons were performed to compare treatment levels for each model 77 .

Growth and necrosis. Initial and final thickness (in Carteriospongia foliascens, Cymbastela coralliophila and
Stylissa flabelliformis), height (in Coscinoderma matthewsi), or sponge tissue depth inside coral cores (in Cliona orientalis), of each sponge (±0.1 mm) was measured with callipers as a 2-dimensional proxy for growth. Three measurements per sponge were averaged to calculate percentage change throughout the experimental and subsequent observational period. Changes in size, partial mortality (i.e. necrosis, assessed through species-specific changes in tissue characteristics, such as changes towards brown/black colour, absence of a healthy pinacoderm and exposure of fibres and skeleton) and loss of photosynthetic symbionts (bleaching) was recorded weekly using a digital camera (Canon S120) with underwater housing and analysed using image analysis software (ImageJ 78 ). Relative growth rates were calculated as the logarithm of the final measure divided by the initial measure, based on both thickness of sponges and their surface area. Percent mortality per tank (n = 3 per treatment) was fitted to nonlinear regression curves using the program Prism v7.01 (GraphPad Software Inc, CA, US). Regression curves were used to calculate the lethal concentrations (LC) of SSC at which 50% (LC 50 ) and 10% (LC 10 ) of the population died. The models were constrained between 0 and 100 with F values set at 50 and 10, for LC 50 and LC 10 , respectively. The curve was tested for normality of the residuals and a replicate test was applied to assess goodness of fit. Symmetric, asymptotic confidence intervals were calculated for the LC values.
Lipid analysis. The concentration of total lipids in sponge tissue was measured over time to assess whether SSC interferes with sponge feeding capability. Samples were analysed from the field controls and at the start of the experiment and at the end of the experimental and observational periods for samples exposed to the 0, 30 and 100 mg L −1 SSCs. Approximately 3 cm 3 of sponge tissue was excised, wrapped in aluminium foil to prevent plasticizer contamination and immediately frozen in liquid nitrogen. Lipids were extracted from approximately 100 mg of freeze-dried ground sample as described in ref. 79 and following modifications in ref. 80, with total lipid content reported as percentage biomass based on a dry weight conversion factor.

Respiration rates.
Changes in sponge respiration rates were measured throughout the experimental and observational period. Samples were dark adapted for a minimum of 30 min before being transferred to 600 mL respiration chambers where they were incubated for 30 min at 27 °C. Constant mixing within the chamber was achieved using a magnetic stir bar and a submergible battery-operated platform. An HDQ30D flexi meter (HACH LDO ™ , CA, US) was used to take three initial readings (i.e. initial O 2 ) and to measure the mg L −1 of O 2 and the % O 2 in each chamber at the end of the incubation. The final oxygen concentration inside the chamber did not drop below 85% saturation in most cases. To control for microbial community respiration, a chamber containing only environmental water (blank) was incubated under identical conditions. Histology. Changes in the structure of the aquiferous system, in particular the number of choanocyte chambers, was examined from histological sections of samples collected at the end of the 28 d experimental period for the 0 and 100 mg L −1 treatments. Samples were fixed for 6 weeks in FAACC fixative (10 mL 40% formaldehyde, 5 mL glacial acetic acid, 1.3 g calcium chloride dehydrate, 85 mL water), and then transferred to 70% ethanol and stored at room temperature. Samples were dehydrated through a graded series of ethanol, embedded in paraffin, sectioned to 5 µm and stained using Mayer's Haematoxylin, and Young's Eosin Erythrosine. Samples were visualized and photographed using an Axioskop 2 plus Microscope and AxioCam MRC5 Digital Camera (Carl Zeiss Microscopy, LLC, US). Choanocyte chambers were only quantified for Cliona orientalis as the presence of spicules in all other species resulted in poor quality sections that precluded accurate choanocyte chamber quantification.
In Cliona orientalis, five fields of view (FOV) per section were examined to quantify the total number of choanocyte chambers at 200×. Choanocyte chambers were identified as circular rings of cells 73 and chamber density was averaged across the 5 FOVs to provide a mean choanocyte chamber density per sponge.
Chlorophyll fluorescence. Changes in photosynthetic capacity (maximum quantum yield) of the sponge's phototrophic symbionts were measured with a Diving-PAM (pulse amplitude modulation) chlorophyll fluorometer (Heinz Walz GmbH, Effeltrich, Germany) for Carteriospongia foliascens, Cymbastela coralliophila and Cliona orientalis as described in ref. 44. Briefly, maximum quantum yield (F v /F m ) 81 measurements were obtained from dark-adapted sponges at weekly intervals throughout the experimental and observational periods. Pigment analysis. Pigment analyses were performed on ~0.2 g of tissue from all phototrophic sponges (Carteriospongia foliascens, Cymbastela coralliophila and Cliona orientalis) at the end of the experimental and recovery periods. Pigments from samples incorporating pinacoderm and mesohyl regions were extracted and analysed as described in ref. 44 and standardized to sponge wet weight. The concentration of Chlorophyll a (hereafter Chl a) was used as a proxy for changes in photosymbiont health/activity (i.e. bleaching) 18 . Differences between Chl a at the end of the 28 d experiment and after the 14 d recovery phase were assessed with a t-test for every treatment and species, separately. Pearson correlations were performed between all the studied pigments. All pigments retrieved by spectrophotometry (i.e. chlorophylls a, b, c, d, Total chlorophylls and carotenoids) were used to build resemblance matrices based on normalized data for each species, separately. Non-metric Multi-Dimensional Scaling (nMDS) plots were created using Euclidean distances. Two factors were determined (i.e. SSC and sampling time, nested to SSC) and examined by PERMANOVA (Permutational multivariate ANOVA based on distances). All multivariate analyses were performed using Primer 6 (Primer-E Ltd, UK).

Microbial community analysis.
Microbial community composition was assessed using Illumina amplicon sequencing of the 16 S rRNA taxonomic marker gene, for field controls, t = 0 controls and at the end of the experimental and observational periods for sponges exposed to 0, 30 and 100 mg L −1 SSC. All samples were immediately frozen in liquid nitrogen and subsequently stored at −80 °C. Water was also collected from each tank at the time of sampling to facilitate a direct comparison with microbes present in the surrounding environment. DNA was extracted from ~0.2 g of sponge tissue using the PowerPlant ® Pro DNA Isolation Kit (MoBio Laboratories, CA, US) according to the Manufacturer's protocol. Microbial communities in seawater were filtered and DNA was extracted as previously described 82 . Sequencing of the 16 S rRNA gene was performed at the Australian Centre for Ecogenomics using primers 515f and 806r and the Illumina HiSeq2500 platform. Sequence data was deposited at the NCBI under the accession number SRP080228.
Amplicon sequence data was processed in Mothur v.1.35.1 83 according to the MiSeq standard operating procedure 84 . Briefly, demultiplexed fastq paired-end reads were first quality-filtered and assembled into contigs (make. contigs and screen.seqs: maxambig = 0, maxhomop = 8, minlength = 100, maxlenght = 292). Aligned reads were reduced to non-redundant sequences and chimeric sequences were detected using Uchime 85 . Aligned sequences were phylogenetically classified based on the Silva reference file v.123, and all unassigned sequences removed (taxon = Chloroplast-Mitochondria-unkown-Eukaryota). Samples with low read numbers were eliminated from the dataset and remaining samples were sub-sampled to 10385 sequences. Pairwise distances were calculated and used for clustering and OTU assignment and OTUs were further classified based on the SILVA v.123 taxonomy.
OTU data was normalised to account for sampling depth and then square-root transformed to reduce the effect of abundant OTUs. Bray-Curtis distance matrices were constructed and visualised using non-metric multidimensional ordinations (nMDS) and principal coordinates analyses (PCO). Permutational analysis of variance (PERMANOVA, using 9999 permutations) was used to determine significant differences in microbial communities. All multidimensional statistical analyses were performed in Primer 6/PERMANOVA. Similarity Percentage (SIMPER) analysis was used to determine the OTUs that contribute to the differences between 2 SSC (0 and 100 mg L −1 ) for each phototrophic species, separately. The 30 OTUs with the most discriminating power from the SIMPER analysis were used to create networks on Cytoscape 3.2.0 86 .