Warming affects predatory faunal impacts upon microbial carbon cycling

1. Ocean warming and the loss of larger (often predatory) fauna are major threats to seabed (benthic) ecosystem functioning. Yet we know little about the combined effects of warming and faunal species loss upon the marine carbon cycle.


Introduction
Present estimates indicate that global temperatures will rise by at least 2 °C by the end of the twenty-first century (IPCC 2014).This will have pronounced impacts upon the species distributions, phenology and the metabolic activity of many organisms, in particular those in marine ecosystems (Pauly et al. 1998;Burrows et al. 2011;Bellard et al. 2012;Poloczanska et al. 2013a;Payne et al. 2016).Through its effects upon the metabolic demands of organisms, ocean warming will alter the strength of interactions between autotrophic (primary producer) and heterotrophic (consumer) organisms and affect the rates and pathways for carbon and nutrient cycling (Daufresne, Lengfellner & Sommer 2009;López-Urrutia et al. 2009;O'Connor et al. 2009).Metabolic theory predicts that the influence of temperature upon organismal metabolism is negatively associated with body-size (Ernest et al. 2003).Consequently, microbial processes that dominate carbon and nutrient cycling in marine sediments are predicted to be particularly sensitive to ocean warming (Gillooly et al. 2001;Weston & Joye 2005;Daufresne et al. 2009;Morán et al. 2015).
Non-vegetated coastal sediments are important to the global carbon cycle, accounting for between 13 and 28 % of total marine carbon burial (~ 226 -418 Tg C yr -1 ; Cai 2011; Bauer et al. 2013;Atwood et al. 2015).Over the past 50 years, their areal coverage has increased by between 25 and 50%, driven by changes in land use, climate change and nutrient enrichment (Atwood et al. 2015).In coastal sediments, the microphytobenthos (MPB) are an assemblage of unicellular eukaryotic algae and cyanobacteria that grow in the upper millimetres of illuminated sediments (MacIntyre, Geider & Miller 1996).As such, the MPB are the main primary producers in non-vegetated sediments, contributing up to 50% of the primary productivity in coastal ecosystems (Underwood & Kronkamp 1999).The MPB play a critical role in carbon fixation and storage in the sediment organic matter pool (Middelburg et al. 2000;Evrard et al. 2010), provide a food source for benthic fauna (Miller, Geider & MacIntyre 1996;Middelburg et al. 2000;Evrard et al. 2010) and, produce extracellular polysaccharides that contribute to sediment stabilisation (Miller et al. 1996;Tolhurst et al. 2006;Spears et al. 2008) and provide a labile organic matter source for heterotrophic microorganisms (Oakes & Eyre 2014).As the area of non-vegetated coastal sediments continues to grow, their responses to the cumulative impacts of faunal species loss and ocean warming will have major implications for carbon cycling in the future ocean.
The impacts of environmental change upon ecosystem processes remains a ecological 'black box' (Burrows et al. 2011;Cardinale et al. 2012;Poloczanska et al. 2013b;Halpern et al. 2015).Stable-isotope labelling experiments provide a powerful tool to open the box, by empirically tracing carbon and energy flow through a biological systems (Middelburg et al. 2000;Boschker, Kromkamp & Middelburg 2005;van Nugteren et al. 2009;Evrard et al. 2010;Hunter et al. 2012;Mayor, Thornton & Zuur 2012;Hunter et al. 2013).By tracing the incorporation and transfer of a carbon source enriched in the rare stable isotope carbon-13 ( 13 C), over a fixed time period, we can test how environmental changes affect various aspects of the carbon cycle, such as carbon fixation and retention (Middelburg et al. 2000;Evrard et al. 2010) or the pathways for organic matter degradation (van Nugteren et al. 2009;van Nugteren, Moodley et al. 2009;Hunter et al. 2012;Mayor et al. 2012;Hunter et al. 2013).In this study, we test how the presence of a benthic predator affected carbon fixation and retention rates in coastal sediments and, whether these putative effects differ between present climatic conditions and predicted future warming (ambient + 2 o C) conditions (IPCC 2014).We focus upon the impacts of the shore crab Carcinas maenas (Linnaeus, 1758), hereafter Carcinus, a globally distributed predator / scavenger, commonly found on both soft and hard substrata in coastal habitats (Crothers 1968).Carcinus is an active bioturbator that excavates soft sediments, transporting buried sediment back to the surface (Queiros et al. 2013).We predicted that the effects of Carcinus upon carbon fixation and retention within coastal sediments will be mediated by ocean warming.We hypothesised that under ambient conditions, physical disturbance of the sediment surface by Carcinus limit the carbon fixation by the microphytobenthos, reducing 13 C-retention within the sediment organic matter pool.Based on previous studies (e.g.Canuel et al. 2007;Spivak et al. 2007;Fanjul et al. 2015) we hypothesised that Carcinus presence would promote the preservation of non-isotopically labelled organic matter within the sediment and limit the accumulation of heterotrophic microbial biomass.As ocean warming is predicted to increase both faunal and microbial metabolic demands, we hypothesised that under future warming conditions (ambient + 2ºC) the direct (and indirect) effects of Carcinus presence upon carbon fixation and retention within coastal sediments would be amplified.

Experimental Design and Set-up
Stable isotope ( 13 C) pulse-chase experiments were conducted during February and March 2015 using outdoor flow-through mesocosms at Queen's University Marine Laboratory (Portaferry, Northern Ireland [Mrowicki & O'Connor 2015]).We assembled 20 mesocosms containing intertidal sediment manipulating two factors: 1. Presence of Carcinus (two levels, present [single crab] and absent); and 2. temperature (two levels, ambient and warming [ambient + 2°C]).Mesocosms were arranged across four outdoor water baths, supplied with a constant flow of sand-filtered water from the adjacent Strangford Lough [flow rate = 27.84(± 4.89) l. h -1 ] and enclosed with plastic mesh lids, with a mesh size of 5 mm, to prevent the crabs from escaping.Elevated temperature treatments were maintained using aquarium heaters (Elite Submersible 300W, Hagen Inc., USA) to warm two of the water baths.Mesocosms received a daily water change, leaving the sediment for 4 hours to simulate low tide.Prior to the daily water change, temperature, salinity, dissolved oxygen and pH were recorded within each mesocosm using a YSI 6-series sonde and multi-meter [YSI Inc.Ohio, USA] (Supplementary Table 1).
Experiments were conducted over 4 weeks (28 days).A single male crab (carapace width = 3.47 [± 0.14] cm, wet biomass = 9.66 [± 2.21] g) placed into five ambient and five warmed mesocosms 7 days after the incubations commenced; a 24 mg (0.1 g.m -2 ) dose of 13 C-labelled sodium bicarbonate was sprayed directly onto the sediment surface of all mesocosms on day 21.The experiments were then incubated for a further 5 days and then destructively sampled to quantify 13 C-incorporation and retention in microbial biomass and the sediment organic matter pool, following Middelburg et al. (2000).
Changes in the biomass and spatial variability of the sediment microphytobenthos (MPB) were measured within each mesocosm using a BBE Moldänke BenthoTorch [BBE Moldänke Gmbh, Schwentinental, Germany] to quantify surficial chlorophyll a (Chla) concentrations after 7 (Week 0), 14 (Week 1), 21 (Week 2) and 28 (Week 3) days (Kahlert & McKie 2014).At each interval five replicate measurements of the surficial Chla were made, within each mesocosm.Mean Chla concentrations within each mesocosm were estimated as a proxy for MPB biomass, with the coefficient of variation of the Chla measurements within each mesocosm providing a metric for MPB patchiness within each mesocosm.
After 28 days, the mesocosm sediments were randomly sampled using four 8 cm diameter push cores.These were horizontally sectioned into 0-0.5;0.5-2; 2-5; 5-10 cm depth fractions.Depth fractions from each of the four cores were pooled, homogenised and lyophilised (-68 °C; 0.0001 mbar), then stored frozen (-20 °C) prior to further analysis (Hunter et al. 2012;Hunter et al. 2013).Total organic carbon (TOC) and the retention of the 13 C-label into the sediment organic carbon pool were determined using a Thermo Flash1112 Elemental Analyser interfaced with a Delta V Isotope Ratio Mass Spectrometer [EA-IRMS].Analytical precision, based on regular replicated analysis of two international standards (USGS 41 and CH7) was < ±0.1 ‰.Prior to analysis, sediment samples were lyophilised, acid fumed with 6 mol l -1 hydrochloric acid to remove inorganic carbon following Hedges and Stern (1984) and dried to constant weight at 60 °C.
Microbial biomass and retention of the 13 C-label within the surficial (0-0.5 cm) sediments were quantified, using polar lipid fatty acids (PLFAs) as biomarkers.PLFAs were extracted from 3 g of lyophilized sediment using a modified Bligh-Dyer extraction protocol (White et al. 1979).Briefly, lipids were extracted over 2 hours using a singlephase extraction mixture of chloroform, methanol and citrate buffer (1:2:0.8v/v/v).Lipid extracts were fractionated on silicic acid columns (6 ml ISOLUTE SIS PE columns, International Sorbent Technologies Ltd, Ystrad Mynach, UK) via sequential elution with chloroform (neutral lipid fatty acids), acetone (glycolipid fatty acids) and methanol (PLFAs).PLFAs were transmethylated to fatty acid methyl esters for analysis.PLFA concentrations and 13 C/ 12 C-ratios were determined on a Thermo GC Trace Ultra gas chromatograph combustion interfaced to a Thermo Delta V Advantage isotope ratio mass spectrometer via a Thermo GC Combustion III (following Thornton et al. 2011).The 13 C-enrichment of the TOC and PLFAs were then determined from the 12 C/ 13 C ratios, following Hunter et al. (2013), correcting for background isotopic ratios of each material (Supplementary tables 2 & 3).All data were scaled to the areal cover (per mesocosm) by converting to values per cm -3 of wet sediment, pooling the data for from each sediment depth fraction (0-0.5;0.5-2; 2-5; 5-10 cm) and normalised to total mesocosm area (0.24 m 2 ).

Data Analysis
We tested for significant interactions and independent effects of Carcinus presence and temperature by analysis of variance, treating Carcinus presence (2 levels: presence, absence) and temperature (2 levels: ambient, ambient + 2 °C) as fixed effects.We included water bath as a random error term within the ANOVA models to account for the potential effects of mesocosm location (Underwood, 1997;Crawley 2007).ANOVA model structure may, thus, be expressed as y ~ C * T + Error (WB), where C = Carcinus presence, T = Temperature and WB is the Water Bath.Data and residuals were visually explored to ensure they met assumptions of normality and homoscedacity (following Zuur, Ieno & Elphick 2010).We tested for the effects of Carcinus presence and temperature upon MPB biomass (sediment chlorophyll a concentrations) and the spatial variability of the MPB community independently during each four weeks of the experiment.The inclusion of week as a covariate would have required the extension of the ANOVA model to include a 3-way interaction term and a temporal autocorrelation structure.This was deemed inappropriate given the sample size within each our treatments (n = 5).
Using the ANOVA model described above, we tested for treatment effects upon sediment TOC and TO 13 C concentrations, total PLFA and 13 C-labelled PLFA concentrations as proxies for microbial biomass and microbial carbon retention at the end of the experimental incubations, and the relative abundance bacterial and microeukaryote fatty acid biomarkers (Supplementary Table 4) within both the total and 13 C-labelled PLFA pools.We tested for statistical significance at p < 0.05, as a reasonable compromise between the risk of false rejection of the null hypothesis (Type I error) and failure to detect a significant effect (Type II error), given the limited replication (n = 5) within each treatment (Underwood, 1997).Where we detect a significant p-value, we calculated the effect size ( 2 ) as the sum of squares for any significant effect(s) / total sum of squares for the ANOVA model, following Cohen (1988).Effect sizes are reported alongside the p-values for each ANOVA in tables 1, 2 and 3. We consider a small effect size where  2 < 0.15; with the associated p-values interpreted with caution (Cohen 1988).Data analysis were conducted in R, using the base and Sciplot packages (R Development Core Team 2009; Morales 2012).

Results
Differences in the concentration and spatial variability of chlorophyll a within surficial sediments were tested independently during each of the four time points (weeks 0, 1, 2, 3 and 4), with Carcinus introduced to the mesocosms during week 1.Carcinus presence resulted in a significant decrease in MPB biomass during weeks 1 and 2 of the experiment, with no effects detected during weeks 0 and 3 (Figure 1a; Table 1a).Our data suggests a significant interaction between Carcinus presence and temperature affected MPB patchiness during week 1 (Fig. 1b; Table 1b), albeit with a small effect size ( 2 =0.112,Table 1b).No significant effects of Carcinus presence and temperature upon MPB patchiness were detected during weeks 2 or 3 (Fig. 1b).We, thus, infer that Carcinus presence affected the growth of the MPB negatively during weeks 1 and 2 of the experiment.No significant effects of water bath, as a nesting factor, were detected upon either MPB biomass or patchiness.
We observed a significant effect of Carcinus presence upon sediment TOC concentrations (Fig. 2a; Table 2a), which decreased by ~ 15 % when crabs were absent. 13C-label retention within the sediment (TO 13 C) accounted for between 0.02 and 0.42 % of the 13 C-labelled bicarbonate dose added to each mesocosm (Fig. 2b; Table 2b).We identified a significant interaction between Carcinus presence and temperature on sediment TO 13 C. Carcinus presence decreased sediment TO 13 C concentrations but only under ambient conditions.Under the future warming conditions, by contrast, TO 13 C concentration within the sediments was greater in the presence of Carcinus.
We used the concentrations of PLFAs in surficial sediments as a proxy for total microbial biomass.Total PLFA concentrations ranged between 11 and 26 mg C mesocosm -1 , equating to between 15 and 36 g microbial biomass within each mesocosm (following Middelburg et al., 2000;Evrard et al., 2010).Within this study, we detected no effects of either Carcinus presence or temperature upon microbial biomass (Fig. 2c; Table 2c).Microbial 13 C-retention was negatively affected by increased temperature (Fig. 2d; Table 2d).
In total, 31 individual PLFA were detected, of which five were identified as bacteriaspecific fatty acids and three polyunsaturated fatty acids (PUFAs) as microeukaryotespecific biomarkers (Supplementary Table 4).Bacterial contributions to the PLFAs exhibited a significant increase in response to warming, with no effects of Carcinus presence (Fig. 3a; Table 3a).Bacterial contributions to the 13 C-labelled PLFAs were affected by significant interactions between Carcinus presence and temperature (Fig. 3b; Table 3b).Carcinus presence reduced 13 C-labelling of the bacterial PLFAs under ambient conditions, and increased 13 C-labelling of the bacterial PLFAs under the warming treatment.
Significant interactions between Carcinus presence and warming affected the relative abundance of both the microeukaryote PUFAs (Fig. 3c; Table 3c) and their contribution to the 13 C-labelled PLFA pool (Fig. 3d; Table 3d) in both cases Carcinus presence and warming resulted in significant decreases in PUFA concentrations and 13C-labelling relative to the ambient-no crab treatment.The most commonly detected polyunsaturated PLFA was 18:26,9, which is widely recognised as a fungal biomarker (Stoeck et al. 2002;Kaiser et al. 2010;Frostegård, Tunlid & Bååth 2011), accounting for between 39.45 (± 1.81) % and 46.02 (± 2.69 %) of the total PUFAs.

Discussion
The study shows that the presence of the mobile epibenthic predator, Carcinus maenas, mediated the retention of newly fixed ( 13 C-labelled) carbon within intertidal sediments, with faunal impacts altered by changes to the ambient temperature.MPB biomass and patchiness, (proxies for assessing the productive potential of coastal sediments, Hicks et al. 2009;Kahlert & McKie 2014), were both initially affected by the introduction of Carcinus to the mesocosms.Where Carcinus were present, the sediment surface was visibly disturbed with track marks, and appeared loss cohesive in texture.Alongside this disturbance the MPB biomass increased and MPB patchiness decreased over the subsequent two weeks, suggesting that faunal reworking of the sediment had little long term effects upon the MPB.This is surprising given that faunal disturbance tends to limit the accumulation of MPB biomass or fresh algal phytodetritus at the sediment surface (Canuel et al. 2007;Spivak et al. 2007;Hicks et al. 2009;Jeffreys et al. 2011;Fanjul et al. 2015).However, benthic diatoms can switch between autotrophy at the sediment surface and heterotrophic fermentation within oxygen and light limited sub-surface sediments (Bourke et al. 2016).We postulate that faunal reworking of the sediments provides a mechanism for the transfer of active MPB cells between the sediment surface and deeper sediment layers, which allows MPB biomass to recover from the initial disturbance event.
Coastal regions are important in the global carbon cycle, where organic matter burial exceeds 226 Tg C yr -1 , of which up to 50 % may be recycled back to the water column as carbon dioxide (Middelburg, Soetaert & Herman 1997;Bauer et al. 2013).Understanding how fauna influence these processes is critical if we are to predict the impacts of the current high rates of biodiversity loss.Within our study, the bulk of the sediment TOC pool consisted of non-labelled organic matter and can, thus, be considered analogous with 'old' carbon stored within the sediment.Carcinus absence was associated with a small, but significant, decrease in the retention of this 'old' carbon within the sediment (Fig. 2a).As an active bioturbator, Carcinus plays a key role reworking coastal sediments (Queiros et al. 2013).Given that sediment and crabs were collected from the same field site, it is reasonable to infer that sediment carbon stocks are in equilibrium with the presence of Carcinus.Consequently, the decrease in TOC associated with the absence of Carcinus suggests that faunal disturbance may limit microbial activity within the mesocosm sediments.This provides further evidence that benthic fauna, such as crabs, play an important role regulating microbial activity and the preservation of organic matter within coastal sediments (Mermillod-Blondin & Rosenberg 2006;Canuel et al. 2007;Spivak et al. 2007;Fanjul et al. 2015).
The 13 C-labelled organic matter represents a sedimentary carbon pool that has been "recently-fixed" by the MPB.This 'recently-fixed' carbon pool is likely to be composed of relatively labile short-chain organic molecules (e.g.lipids, amino acids), which may be preferentially utilised as a carbon source by the sediment microbial community (e.g.van Nugteren et al. 2009;Miyatake et al. 2014).Our study identified contrasting effects of Carcinus presence upon the concentrations of this 'recently-fixed' carbon under both the present and future warming conditions (Fig. 2b).No significant differences in surficial MPB biomass were observed during the isotope-labelling phase of the experiment (Fig. 1a) and, under ambient conditions, Carcinus presence had no effect upon 13 C-fixation into the organic matter pool (Fig. 2B) Consequently, crab activity had no effect upon Cfixation by the MPB contrasting with previous studies that have identified faunal disturbance of the sediment surface as an important control upon microbial carbon fixation (Mermillod-Blondin & Rosenberg 2006).Under the predicted future-warming treatment, however, the absence of Carcinus is associated with reduced retention of 13 Clabelled carbon in the sediment.This indicates that under the predicted warming conditions tested, increased metabolic activity by the sediment microbial community, leading to more rapid fixation of the 13 C-label by the MPB and more rapid mineralisation of 'recently-fixed' carbon by heterotrophic microorganisms (Gillooly et al. 2001;Weston & Joye 2005;Daufresne et al. 2009;Morán et al. 2015).
The microbial response to Carcinus presence and warming were tested in the surficial sediment layer (0-0.5 cm), using PLFAs as biomarkers.PLFAs provide a powerful tool in this context, allowing both microbial biomass and incorporation of the 13 C-label to be quantified (Middelburg et al. 2000;Mayor et al. 2012;Hunter et al. 2013).Microbial biomass (estimated from PLFA concentrations) was unaffected by either Carcinus presence or temperature (Fig. 2c), however, we observed a significant decrease in 13 Clabel retention within the PLFAs (Fig 2d).This is likely to reflect increased metabolic activity within the microbial community as a response to ocean warming (Weston & Joye 2005).Carcinus presence played no significant role in determining 13 C-label retention within the PLFAs.We suggest that turnover of microbial biomass at the sediment surface (reviewed in Mermillod-Blondin & Rosenberg 2006), combined with temperature-dependent increases in microbial metabolism (Weston & Joye 2005;Arndt et al. 2013) may explain this observation.
Our results show that faunal species loss has pronounced effects upon microbial community responses to ocean warming.In this context, Carcinus presence had a significant influence upon both the structure of the microbial community and its active component, as revealed from the relative abundance of bacteria-specific FAs and PUFAs within the total PLFA and 13 C-labelled PLFA profiles.PLFAs can provide useful information on microbial community structure in aquatic sediments (e.g.Stoeck et al. 2002;Boschker et al. 2005;Mayor et al. 2012), albeit with caveats regarding their limited taxonomic resolution (Frostegård et al. 2011).We observed broad trends in the response of the microbial community within our mesocosms.This is characterised by temperature-driven decreases in the bacterial FAs, which reflect the impacts of ocean warming on microbial cell-size (Morán et al. 2015).There was a strong negative effect of crab presence upon the PUFAs under ambient conditions.PUFAs can be considered both as a biomarker of microeukaryote biomass (Stoeck et al. 2002;Kaiser et al. 2010;Frostegård et al. 2011) and an indicator of the relative lability of the sediment organic matter pool (Canuel et al. 2007;Spivak et al. 2007).Within our study, PUFAs were dominated by a single fungal biomarker (18:27) suggesting that, under ambient conditions, Carcinus may suppress the accumulation of fungal biomass within surficial sediments.By contrast, PUFA concentrations were lower under future warming conditions, perhaps as a consequence of increases in resource competition and biomass turnover within the microbial community (Daufresne et al. 2009;Sarmento et al. 2010).

Accepted Article
This article is protected by copyright.All rights reserved.
The active microbial community, as revealed by 13 C-labelling of the PLFAs, were sensitive to changes in Carcinus presence under both ambient and future-warming conditions.Both bacterial and microeukaryote biomarkers exhibited similar responses, with Carcinus presence associated with decreased contribution to the active microbial community under ambient conditions, but increasing their contributions under warming conditions.Our study builds upon previous studies which demonstrate that crabs play a key role in coastal sediments, controlling accumulation of microbial biomass and the preservation of organic matter (Canuel et al. 2007;Spivak et al. 2007;Fanjul et al. 2015).This reinforces the concept of the benthic fauna as a 'gearbox' that regulates microbial activity in marine sediments (van Nugteren et al. 2009;Hunter et al. 2012) and support the observation that the responses of faunal communities to rising ocean temperature will have cascading effects upon benthic primary producers (O'Connor et al. 2009;Mrowicki & O'Connor 2015).
Retention of the 13 C-label within the mesocosm sediment was low compared with other studies that investigated label uptake by the MPB (e.g.Middelburg et al. 2000;Boschker et al. 2005;van Oevelen et al. 2006).This reflects the relatively low dose of 13 C-labelled sodium bicarbonate introduced into each mesocosm, and potential losses associated with washout of unused 13 C-labelled sodium bicarbonate; mineralisation of 13 C-labelled organic matter and flux of dissolved organic carbon between the sediment and overlying water (e.g.Middelburg et al. 2000;Evrard et al. 2010).Given that coastal sediments are typically respiration-dominated systems (Woulds et al. 2009;Woulds et al. 2016), mineralisation of 13 C-labelled organic matter is likely to represent a major un-quantified process within our study.Caution is therefore advised in comparing these data with other sediment carbon budgets, derived using similar methodologies (Middelburg et al. 2000;Evrard et al. 2010;Woulds et al. 2016).Our study builds upon observations of cascading effects of crab predation upon sediment organic matter composition in coastal sediments (Canuel et al. 2007;Spivak et al. 2007;Fanjul et al. 2015), and provides a direct test of how faunal species loss affects sediment carbon sequestration under both present climatic conditions and predicted future-warming conditions.
Our study demonstrates that both faunal presence and temperature are important regulators for organic matter retention within coastal sediments.Faunal presence was the primary driver of changes in 'old' organic carbon within the mesocosm, whilst strong interactions between faunal presence and temperature determine the fate of 'recently-fixed' organic matter.Bulk analysis of the TOC and total PLFAs was, however, limited in its scope to identify the interacting effects of Carcinus presence and temperature.Compound-specific analysis of the PLFA profiles, allowed us, to discern how faunal presence stimulated 13 C-incorporation by heterotrophic microbes under the predicted-future warming treatment, and suppressed the accumulation of microeukaryote (fungal) biomass under ambient conditions.These effects are clearly driven by changes in faunal bioturbation and the rate-limiting effects of temperature upon microbial activity.
We acknowledge that the results of mesocosm experiments are largely illustrative, and cannot easily be scaled to natural systems (Oviatt 1994;Carpenter 1996;Queiros et al. 2015).If we are to mitigate the impacts of global biodiversity loss, however, manipulative experiments are an important tool to test potential impacts on ecosystem processes (Duffy 2009;Cardinale et al. 2012;Stewart et al. 2013;Donohue et al. 2017).Previous mesocosm studies highlight the importance of epibenthic predators, such as crabs, as mediators of organic matter preservation in coastal sediments (Canuel et al. 2007;Spivak et al. 2007;Atwood et al. 2015;Donohue et al. 2017;Fanjul et al. 2015).
Here we demonstrate that under future warming conditions, epibenthic predator loss leads to an increase in the relative importance of microbial C cycling pathways in coastal sediments.Whilst we cannot easily predict ecosystem-scale responses, this may result in decreasing carbon storage within coastal sediments.Consequently the combined effects of predator loss and ocean warming are likely to have adverse effects upon a range of coastal ecosystem services that are relevant for climate regulation, waste processing, flood protection and support of fisheries (reviewed in Lopes & Vidiera 2013;Isbell et al. 2017).

Accepted Article
This article is protected by copyright.All rights reserved.

Accepted Article
This article is protected by copyright.All rights reserved.

Accepted Article
This article is protected by copyright.All rights reserved.

FiguresFigure 1 .
Figures Figure 1.Effects of Carcinus presence and temperature upon the microphytobenthos.Temporal changes in mean (± standard error) a) MPB Biomass (surficial chlorophyll a concentrations) and b) spatial variation in the MPB assemblage between Carcinus present (white ) and absent (grey) treatments, under both ambient ( ) and warming ( ) temperature treatments.Inserts show pooled means (± standard error) where significant independent effects of either Carcinus presence (white) / absence (grey) or temperature (labelled) where detected.Significance levels: *** p < 0.001; ** p < 0.01; * p < 0.05.Where significant interactions were identified, groups labeled with the same lowercase letter are not significantly different (p > 0.05; Tukey's tests).

Figure 2 .
Figure 2. Effects of Carcinus presence and temperature upon the sediment geochemistry.Mean (± standard error) concentrations of a) Total Organic Carbon b) 13 C-labelled Organic Carbon (TO 13 C) c) Total PLFAs and d) 13 C-labelled PLFAs between Carcinus present (white ) and absent (grey) treatments, under both ambient ( ) and warming ( ) temperature treatments.Inserts show pooled means (± standard error) where significant independent effects of either Carcinus presence (white) / absence (grey) or temperature (labelled) where detected.Significance levels: *** p < 0.001; ** p < 0.01; * p < 0.05.Where significant interactions were identified, groups labeled with the same lowercase letter are not significantly different (p > 0.05; Tukey's tests).

Figure 3 .
Figure 3. Effects of Carcinus presence and temperature upon the sediment microbial community.Mean (± standard error) contributions of a) bacterial fatty acids b) 13 C-labelled bacterial Fatty Acids, c) microeukaryote PUFAs and d) 13 C-labelled microeukaryote PUFAs [relative to the total PLFA or 13 C-labelled PLFA pools] between Carcinus present (white ) and absent (grey) treatments, under both ambient ( ) and warming ( ) temperature treatments.Inserts show pooled means (± standard error) where significant independent effects of either Carcinus presence (white) / absence (grey) or temperature (labelled) where detected.Significance levels: *** p < 0.001; ** p < 0.01; * p < 0.05.Where significant interactions were identified, groups labeled with the same lowercase letter are not significantly different (p > 0.05; Tukey's tests).

Table 1 .
ANOVA summary tables of the effects of Carcinus

Table 1 . a) MPB Biomass (Chlorophyll a)
This article is protected by copyright.All rights reserved.b)MPB Patchiness