E ﬀ ects of the Bioturbating Marine Yabby Trypaea australiensis on Sediment Properties in Sandy Sediments Receiving Mangrove Leaf Litter

: Laboratory mesocosm incubations were undertakento investigate the inﬂuenceof burrowing shrimp Trypaea australiensis (marine yabby) on sediment reworking, physical and chemical sediment characteristics and nutrients in sandy sediments receiving mangrove ( Avicennia marina ) leaf litter. Mesocosms of sieved, natural T . australiensis inhabited sands, were continually ﬂushed with fresh seawater and pre-incubated for 17 days prior to triplicates being assigned to one of four treatments; sandy sediment (S), sediment + yabbies (S + Y), sediment + leaf litter (organic matter; S + OM) and sediment + yabbies + leaf litter (S + Y + OM) and maintained for 55 days. Mangrove leaf litter was added daily to treatments S + OM and S + Y + OM. Luminophores were added to mesocosms to quantify sediment reworking. Sediment samples were collected after the pre-incubation period from a set of triplicate mesocosms to establish initial conditions prior to the imposition of the treatments and from the treatment mesocosms at the conclusion of the 55-day incubation period. Yabbies demonstrated a clear e ﬀ ect on sediment topography and leaf litter burial through burrow creation and maintenance, creating mounds on the sediment surface ranging in diameter from 3.4 to 12 cm. Within S + Y + OM sediments leaf litter was consistently removed from the surface to sub-surface layers with only 7.5% ± 3.6% of the total mass of leaf detritus added to the mesocosms remaining at the surface at the end of the 55-day incubation period. Yabbies signiﬁcantly decreased sediment wet-bulk density and increased porosity. Additionally, T . australiensis signiﬁcantly reduced sediment bio-available ammonium (NH 4 + bio ) concentrations and altered the shape of the concentration depth proﬁle in comparison to the non-bioturbated mesocosms, indicating inﬂuences on nutrient cycling and sediment-water ﬂuxes. No signiﬁcant changes for mean apparent biodi ﬀ usion coe ﬃ cients ( D b ) and mean biotransport coe ﬃ cients ( r ), were found between the bioturbated S + Y and S + Y + OM mesocosms. The ﬁndings of this study provide further evidence that T . australiensis is a key-species in shallow intertidal systems playing an important role as an ‘ecosystem engineer’ in soft-bottom habitats by signiﬁcantly altering physical and chemical structures and biogeochemical function.

Abstract: Laboratory mesocosm incubations were undertaken to investigate the influence of burrowing shrimp Trypaea australiensis (marine yabby) on sediment reworking, physical and chemical sediment characteristics and nutrients in sandy sediments receiving mangrove (Avicennia marina) leaf litter. Mesocosms of sieved, natural T. australiensis inhabited sands, were continually flushed with fresh seawater and pre-incubated for 17 days prior to triplicates being assigned to one of four treatments; sandy sediment (S), sediment + yabbies (S+Y), sediment + leaf litter (organic matter; S+OM) and sediment + yabbies + leaf litter (S+Y+OM) and maintained for 55 days. Mangrove leaf litter was added daily to treatments S+OM and S+Y+OM. Luminophores were added to mesocosms to quantify sediment reworking. Sediment samples were collected after the pre-incubation period from a set of triplicate mesocosms to establish initial conditions prior to the imposition of the treatments and from the treatment mesocosms at the conclusion of the 55-day incubation period. Yabbies demonstrated a clear effect on sediment topography and leaf litter burial through burrow creation and maintenance, creating mounds on the sediment surface ranging in diameter from 3.4 to 12 cm. Within S+Y+OM sediments leaf litter was consistently removed from the surface to sub-surface layers with only 7.5% ± 3.6% of the total mass of leaf detritus added to the mesocosms remaining at the surface at the end of the 55-day incubation period. Yabbies significantly decreased sediment wet-bulk density and increased porosity. Additionally, T. australiensis significantly reduced sediment bio-available ammonium (NH 4 + bio ) concentrations and altered the shape of the concentration depth profile in comparison to the non-bioturbated mesocosms, indicating influences on nutrient cycling and sediment-water fluxes. No significant changes for mean apparent biodiffusion coefficients (D b ) and mean biotransport coefficients (r), were found between the bioturbated S+Y and S+Y+OM mesocosms. The findings of this study provide further evidence that T. australiensis is a key-species in shallow intertidal systems playing an important role as an 'ecosystem engineer' in soft-bottom habitats by significantly altering physical and chemical structures and biogeochemical function.

Introduction
Shallow intertidal habitats typically support a high biomass of benthic assemblages [1] and provide a range of important ecosystem services [2,3]. Understanding factors and processes that regulate these services and assemblages is therefore important. Mangrove trees are characteristic of subtropical and tropical intertidal systems, and while organic matter pools within these ecosystems are often the net result of different inputs [4,5], mangrove leaf litter provides an important source of organic carbon and nutrients [6,7].
In shallow intertidal systems, bioturbation influences sediment characteristics (e.g., particle size distributions, permeability, microbiology and nutrient concentrations [8][9][10][11][12]) altering biogeochemical regimes for organic matter degradation. Sediment reworking processes increase the rate of the burial of newly deposited organic matter, stimulating mineralisation and respiration processes [13][14][15]. Additionally, infauna ventilation along with benthic flora introduce dissolved oxygen to deeper sediment strata creating shifting mosaics of sediment redox conditions [16][17][18][19]. Redox cycles and oscillations within bioturbated sediments are proposed to favor more rapid and complete decomposition of organic detritus than is possible under constant conditions or unidirectional redox changes [20,21]. Infauna plays an important role in the cycling of organic material often having significant effects on the structure and biogeochemistry of sediments [13,14,21,22].
Deposit feeders influence organic matter mineralisation and nutrient regeneration rates predominantly through the physical breakdown of particulate matter, which increases the total surface area available for bacterial colonisation, and through the mixing of freshly deposited or harvested particulate organic matter into the deeper sediment strata [10,14]. These organisms may inhabit permanent or non-permanent burrows, feeding upon surface and sub-surface sediments [23,24], which are subsequently either transported to the sediment surface or to deeper anoxic sediments as faecal castings [25,26].
This study investigated the effect of the marine yabby T. australiensis on sediment reworking, physical and chemical sediment characteristics, and nutrient distribution in sandy sediments receiving mangrove (Avicennia marina) leaf litter in an orthogonal mesocosm experiment. Trypaea australiensis and A. marina are both important case study species because of their widespread co-existence along east-coast Australia [33][34][35][36] and the potential for infauna to alter physical and chemical sediment conditions, including sediment topography, nutrient concentrations and nutrient exchanges, organic matter mineralisation and nutrient cycling in sandy sediments receiving leaf litter. It was hypothesised that in concert with the bioturbating infauna the addition of mangrove leaf litter would have implications on nutrient cycling, organic matter mineralisation and nutrient concentrations. This study provides insights into typical ecosystem services provided by these widespread intertidal habitats.

Experimental Set-up and Design
Sediment, yabbies (T. australiensis) and mangrove (A. marina) leaf litter were collected during 2007 from the Gold Coast Broadwater (Australia) and maintained as described in Dunn et al. [32]. The study region is characterised by bioturbating infauna including T. australiensis, which is one of the dominant macrofauna species [26]. Additionally, A. marina stands fringe large areas of the northern zone of the Gold Coast Broadwater and adjoining southern Moreton Bay.
Sediments retrieved from the collection site were sandy in nature with 99% >63 µm [30]. Sediments were sieved, homogenised and transferred to 15 mesocosms (75 cm deep × 36 cm internal diameter) to a sediment depth of 58 ± 2.5 cm. Mesocosm dimensions were selected to be both practical and accommodate the reported burrow characteristic of T. australiensis [23]. Sediment filled mesocosms were left to stabilise (17 days) to permit sediment physicochemical gradients to re-establish. During the equilibration and incubation period mesocosms were maintained at 22 ± 2 • C, and continually flushed with seawater from the Broadwater at a rate of~15 L d −1 using irrigation tubing connected to a 20,000 L reservoir which contained sufficient water for the entire study. As T. australiensis commonly occur in sub-tidal zones a constant water depth of 17 ± 2.5 cm was maintained within each mesocosm over the duration of the study. Mesocosms were aerated with aquarium air-stones, which along with a small aquarium pump ensured gentle water circulation (Figure 1), whilst avoiding sediment resuspension. At the conclusion of the stabilisation period three mesocosms (S initial ) were sampled to determine initial sediment conditions before triplicate mesocosms were assigned to four treatments, control-sediment only (S), sediment + yabbies (S+Y), sediment + leaf litter (organic matter; S+OM) and sediment + yabbies + leaf litter (S+Y+OM; Figure 1). On day zero of the incubation, eight T. australiensis (80 ind. m −2 ) were added to each S+Y (ind. mean weight 4.13 ± 0.37 g, n = 24) and S+Y+OM (ind. mean weight 4.27 ± 0.87 g, n = 24) mesocosms. Records of the behaviour and surface sediment features produced by the burrowing and feeding activities of the yabbies were routinely recorded. On day three after all yabbies had formed burrows, OM additions (0.125 g wet wt day −1 A. marina leaf litter with a mean C content of 39% dry wt. and C:N ratio of 36.9 ± 2.4 [32]) were initiated in S+OM and S+Y+OM treatments and were repeated daily until the completion of the incubations. Leaf litter additions were equivalent to 4.5 t ha −1 y −1 , which corresponds to the average annual leaf fall for A. marina within Moreton Bay [37]. Additionally, on day three natural sediments coated with a fluorescent dye (luminophores, green colour, 30 g, mode 260 µm, Partrac Ltd., Heathfield, UK) were evenly deposited on the sediment surface of all mesocosms in accordance with Hedman et al. [38]. The size of the coated sediment tracers corresponds to the mean grain size previously determined at the sediment collection site [30].
Mesocosms were incubated for 55 days under the same conditions described for the stabilisation period before final sediment conditions were assessed to determine the effects of T. australiensis on sediments receiving mangrove leaf litter. The overall study timeline is shown in Figure 2.    Schematic representation of the overall study timeline (Note: dx represents study day). Figure 2. Schematic representation of the overall study timeline (Note: d x represents study day).
Prior to the collection of sediment cores, residual leaf detritus remaining on the sediment surface was carefully collected, dried and weighed.

Sediment Reworking Quantification
Sediment sub-samples assigned for tracer retrieval (~25 g) were dried at 60 • C. Luminophores from each depth horizon were counted and weighed and expressed as a percentage of total retrieved luminophores within each mesocosm.
Quantification of sediment reworking was then based on the tracer distribution data using the gallery-diffusion model of macrofaunal reworking (e.g., Hedman et al. [38]; François et al. [40]). This model allows for a description of both the diffusive-like mixing of particles in the region of intense burrowing activity (D b : apparent biodiffusion coefficient) and the rapid transport of organic and inorganic material from the upper sediment layers to lower regions of reworking (r: non-local biotransport coefficient).

Analytical Techniques
Porosity was determined directly from sediment wet-bulk density and water content values [41]. Sediment organic matter content was determined by the loss-on-ignition (LOI 550 ) method, as outlined by Heiri et al. [42]. Sediment C, N, δ 13 C and δ 15 N samples were pretreated as described in [5] and analysed using an elemental analyzer (EA3000, Eurovector, Pavia, Italy) connected to a mass spectrometer (Isoprime, GV Instruments, Wythenshawe, UK) at the Stable Isotope Laboratory (Griffith University, Nathan Campus). Sediment NH 4 + bio concentrations were determined from aliquots of the homogenised sediment in 2 M KCl. Following extraction, samples were centrifuged and the supernatant filtered (GF/F, Millipore, USA) before NH 4 + concentrations were determined using an automated flow injection colourimetric analyser (Easychem Plus Random Access Analyzer; Systea Analytical Technologies, Italy) at the Griffith School of Environment, Environmental Chemistry Laboratory (Griffith University, Gold Coast Campus). Due to the freezing of the samples, the measured NH 4 + pool would include ammonium released from bacterial cells due to membrane damage during the freezing process [43,44].

Statistical Analysis
Select parameters for each treatment and depth are presented as mean values ± one standard deviation (n = 3). A comparison of the day 55 sediment conditions was analysed by two-way ANOVA with two fixed factors (treatment and depth) and the interaction term, to test if treatments contributed to measured differences among sediment parameters measured, and/or with sediment depth. Initial exploratory analyses revealed that the dependent variables required log (x + 1) transformation to ensure that linearity, homogeneity of variance and normality assumptions were satisfied. Criteria of P < 0.01 was used to determine significant differences using SPSS for Windows (SPSS Inc, USA, version 22).
Non-metric multidimensional scaling (NMDS) was used to ordinate sediment treatments from depth matrices by firstly normalising the data to account for different scales of measure in the data before the data were log + 1 transformed to assist with reducing outliers [45]. As part of the NMDS analysis a percentage correlation cut-off of 90% was used in the BIOENV routine. To examine the hypothesis that sediment parameter concentrations/values differed with depth, and among treatments, datasets were tested using PERMANOVA on a Euclidean dissimilarity measure with the terms sediment depth and treatment both fixed, and an interaction term (treatment × depth) [46] using PRIMER (version 6).

Visual Observations
Non-bioturbated mesocosms (i.e., S and S+OM) maintained a smooth uniform surface throughout the study with deposited green tracers still visible (Figure 3a). Immediately upon introduction to the mesocosms (i.e., S+Y and S+Y+OM treatments), T. australiensis commenced burrow construction, where sediments brought to the surface formed mounds (Figure 3b,c, respectively). Mounds typically had one to two openings, ranging in size during the study period from 0.2 to 1.3 cm diameter. Burrow openings were observed to open and close on a daily basis, whilst also changing position within the mounds. Additionally, mound dimensions and positioning shifted throughout the study. Burrow mound diameters varied within, and between, mesocosms with minimum and maximum recorded diameters of 3.4 cm and 12 cm, respectively. Burrow mounds also demonstrated instances of collapse, resulting in funnelling at the sediment surface, producing downward transport of surface sediment. Overall, the presence of T. australiensis resulted in a complex topography of mounds and depressions and the disappearance of the green tracers from the surface sediments (Figure 3b,c). Following initial burrow construction, yabbies were rarely observed at the sediment surface ( Figure 3b).
Dark grey sediments were routinely observed within the burrow mounds (Figure 3c), indicating the transport of anoxic sediment containing black iron monosulfide from depth. These mounds retained dark colouration for a matter of hours, before becoming oxidised to a colour similar to the surface sediment.
Black sulfidic layers were identified around, and between, leaf litter and the sediment surface (S+OM and S+Y+OM). Periodically, white bacterial growth was also observed around and on the upper surface of leaf fragments (Figure 3d), indicating colonisation by filamentous sulfur oxidising chemoautotrophs. The presence of the filamentous sulfur oxidising bacteria indicates that the detritus created localised zones of anoxia where organic matter degradation was mediated through sulfate reduction.
Within bioturbated S+Y+OM mesocosms, most of the added leaf litter rapidly disappeared from the sediment surface, as it was either buried below ejected sediment within the burrow mounds or subducted into the burrows. Leaf litter remaining on the sediment surface on day 55 accounted for 97.9% ± 0.1% and 7.5% ± 3.6% of the total mass of leaf litter added to the S+OM and S+Y+OM mesocosms over the study period, respectively. During sediment slicing, fragments of buried leaf matter surrounded by zones of black sulfidic sediment were clearly visible within the bioturbated S+Y+OM mesocosms.  [17,18] protruding from the sediment in (a-c) (data not presented), small aquarium pump and luminophore 'tracers', sediments coated with a green fluorescent dye, used during the study.

Tracer Profiles
Luminophore tracers were not transported downwards into sub-surface sediments (i.e., buried) in the absence of T. australiensis during mesocosm maintenance, incubations and sampling. Within non-bioturbated mesocosm treatments 100% of the recovered luminophores were found on the sediment surface within the first sampled sediment horizon (0-1 cm, data not shown). Tracer distributions in the bioturbated mesocosms exhibited an initial exponential decrease in the percentage of retrieved luminophores with sediment depth (Figure 4). Within S+Y and S+Y+OM mesocosms, luminophores were present to depths of 40 to 55 cm indicating sediment reworking. Within the bioturbated sediments, 76.43% ± 5.89% and 79.73% ± 10.77% of total luminophores were recovered within the upper 5 cm of sediment in the S+Y and S+Y+OM mesocosms, respectively (Figure 4a (S+Y) and 4b (S+Y+OM)). No significant difference in the percentage of luminophores retrieved was observed between the S+Y and S+Y+OM mesocosms for any depth horizon.  [17,18] protruding from the sediment in (a-c) (data not presented), small aquarium pump and luminophore 'tracers', sediments coated with a green fluorescent dye, used during the study.

Tracer Profiles
Luminophore tracers were not transported downwards into sub-surface sediments (i.e., buried) in the absence of T. australiensis during mesocosm maintenance, incubations and sampling. Within non-bioturbated mesocosm treatments 100% of the recovered luminophores were found on the sediment surface within the first sampled sediment horizon (0-1 cm, data not shown). Tracer distributions in the bioturbated mesocosms exhibited an initial exponential decrease in the percentage of retrieved luminophores with sediment depth (Figure 4). Within S+Y and S+Y+OM mesocosms, luminophores were present to depths of 40 to 55 cm indicating sediment reworking. Within the bioturbated sediments, 76.43% ± 5.89% and 79.73% ± 10.77% of total luminophores were recovered within the upper 5 cm of sediment in the S+Y and S+Y+OM mesocosms, respectively (Figure 4a (S+Y) and Figure 4b (S+Y+OM)). No significant difference in the percentage of luminophores retrieved was observed between the S+Y and S+Y+OM mesocosms for any depth horizon.

Sediment Reworking Coefficients
T. australiensis sediment reworking coefficients provided by the gallery-diffusion model are shown in Table 1. The mean apparent biodiffusion coefficients induced by macrofaunal sediment reworking (Db), were 6.4 ± 1.9 cm 2 y −1 and 5.9 ± 2.1 cm 2 y −1 for S+Y and S+Y+OM mesocosms, respectively. Whereas, mean biotransport coefficients (r) were 3.7 ± 2.1 y −1 and 2.8 ± 2.9 y −1 for S+Y and S+Y+OM mesocosms, respectively. No significant differences were observed between the bioturbated S+Y and S+Y+OM mesocosms for either the diffusion or biotransport coefficients.

Sediment Reworking Coefficients
T. australiensis sediment reworking coefficients provided by the gallery-diffusion model are shown in Table 1. The mean apparent biodiffusion coefficients induced by macrofaunal sediment reworking (D b ), were 6.4 ± 1.9 cm 2 y −1 and 5.9 ± 2.1 cm 2 y −1 for S+Y and S+Y+OM mesocosms, respectively. Whereas, mean biotransport coefficients (r) were 3.7 ± 2.1 y −1 and 2.8 ± 2.9 y −1 for S+Y and S+Y+OM mesocosms, respectively. No significant differences were observed between the bioturbated S+Y and S+Y+OM mesocosms for either the diffusion or biotransport coefficients. Table 1. Calculated apparent biodiffusion coefficients (D b ) and biotransport coefficients (r) determined from the gallery-diffusion model (François et al. [40]) for the bioturbated sediment + yabbies (S+Y) and sediment + yabbies + leaf litter (S+Y+OM) mesocosms.

Physical and Chemical Sediment Profiles
Profiles of sediment wet-bulk density in the non-bioturbated S and S+OM mesocosms on day 55 were similar to those of the initial sediment (Figure 5a). Sediment bulk density was significantly influenced by the presence of T. australiensis (two-way ANOVA, P = 0.001, Table 2). Sediment wet-bulk density values ranged between 1.51 and 2.17 g cm −3 with depth-averaged values in the bioturbated treatments (1.81 ± 0.11 g cm −3 ) being significantly lower than the non-bioturbated sediments (1.95 ± 0.09 g cm −3 ). Similarly, there was a significant difference in sediment porosity between treatments (two-way ANOVA, P = 0.001, Table 2) with sediments in the bioturbated treatments having higher porosity than those in the non-bioturbated treatments (Supplementary File 1). Additionally, there was a significant trend of decreasing porosity with depth across treatments ( Table 2). Profiles (mean ± standard deviation) of (a) wet-bulk density and (b) organic C in each mesocosm treatment: control-sediment only (S), sediment + yabbies (S+Y), sediment + leaf litter (organic matter; S+OM), sediment + yabbies + leaf litter (S+Y+OM) treatments, and initial sediment conditions (Sinitial). The sediment C:N ratio was higher in the 0-20 cm section of the depth profile in the S+Y+OM treatment, compared to all other treatments (Figure 6a). No significant differences were observed in sediment δ 13 C or δ 15 N values between treatments (Table 2) with depth-averaged δ 13 C values ranging from -24.54‰ ± 1.24‰ (S+Y+OM) to -23.71‰ ± 2.12‰ (S+Y; Figure 6b) and depth-averaged δ 15 N values ranging from 6.97‰ ± 2.44‰ (S+Y) to 7.94‰ ± 2.69 ‰ (S; Supplementary File 3). However, there was a significant depth effect for both sediment δ 13 C (2-way ANOVA, P = 0.005) and δ 15 N (two-way ANOVA, P = 0.001) profiles (Table 2). Mean δ 13 C values tended to increase in the deeper sediments, while δ 15 N values decreased with sediment depth.
Depth profiles of NH4 + bio were significantly influenced by the presence of T. australiensis (two-way ANOVA, P = 0.003), with lower NH4 + bio concentrations recorded throughout the 0-55 cm depth profiles in the bioturbated mesocosms (Figure 7). Depth averaged NH4 + bio concentrations within the bioturbated S+Y and S+Y+OM treatments were 32.3 ± 20.3 nmol g −1 dry wt. and 47.8 ± 49.0 nmol g −1 dry wt., respectively. In comparison, concentrations in the non-bioturbated S and S+OM treatments were 84.4 ± 54.0 nmol g −1 dry wt. and 81.0 ± 62.1 nmol g −1 dry wt., respectively. Furthermore, NH4 + bio concentrations were more constant with depth in the bioturbated sediments with little change observed at depths below 5 cm (Figure 7), whereas, in the non-bioturbated sediments NH4 + bio concentrations increased with sediment depth to a depth of 20-30 cm. A significant depth effect was observed across treatments (two-way ANOVA, P = 0.001, Table 2).

Figure 5.
Profiles (mean ± standard deviation) of (a) wet-bulk density and (b) organic C in each mesocosm treatment: control-sediment only (S), sediment + yabbies (S+Y), sediment + leaf litter (organic matter; S+OM), sediment + yabbies + leaf litter (S+Y+OM) treatments, and initial sediment conditions (S initial ). Sediment organic matter content (organic C % and LOI 550 ) profiles tended to be greater in the upper horizons (Figure 5b and Supplementary File 2), however, no clear trends or significant differences with depth or treatment were observed ( Table 2). In contrast the C:N ratio of the sediment demonstrated significant differences between treatments (two-way ANOVA, P = 0.001, Table 2) with depth-integrated average ratios ranging between 17.12 ± 2.86 and 22.69 ± 7.34 in the S and S+Y+OM treatments, respectively. Sediment organic C content values ranged between 0.05% and 0.43% with depth averaged values ranging between 0.06% ± 0.02% (S) and 0.14% ± 0.09% (S+Y+OM). Additionally, measured LOI 550 values varied between 0.04% and 1.64% with depth averaged values ranging from 0.39% ± 0.16% (S+Y+OM) to 0.52% ± 0.47% (S+OM). The sediment C:N ratio was higher in the 0-20 cm section of the depth profile in the S+Y+OM treatment, compared to all other treatments (Figure 6a). No significant differences were observed in sediment δ 13 C or δ 15 N values between treatments (Table 2) with depth-averaged δ 13 C values ranging from -24.54% ± 1.24% (S+Y+OM) to -23.71% ± 2.12% (S+Y; Figure 6b) and depth-averaged δ 15 N values ranging from 6.97% ± 2.44% (S+Y) to 7.94% ± 2.69 % (S; Supplementary File 3). However, there was a significant depth effect for both sediment δ 13 C (2-way ANOVA, P = 0.005) and δ 15 N (two-way ANOVA, P = 0.001) profiles (Table 2)  concentrations were more constant with depth in the bioturbated sediments with little change observed at depths below 5 cm (Figure 7), whereas, in the non-bioturbated sediments NH 4 + bio concentrations increased with sediment depth to a depth of 20-30 cm. A significant depth effect was observed across treatments (two-way ANOVA, P = 0.001, Table 2).

Relevance of Experimental Approach and Observed Behaviour
Mesocosms and microcosms have commonly been used to investigate the influences of burrowing macrofauna on sediment reworking, physicochemical properties, nutrient fluxes and
Upon being added to the sediment during this study, T. australiensis immediately started burrowing (Section 3.1), likely due to their instinctive behaviour to seek refuge from predation whilst exposed on the sediment surface [62]. After the initial burrow excavation, phase T. australiensis were only seen on the sediment surface when dumping excavated burrow sediments. Burrow mounds were formed in less than one day following yabby introduction, while sediment expulsion from the burrows continued throughout the study, creating a shifting mosaic of mound distributions and sizes. Mound diameters observed during this study correspond to those reported by Webb and Eyre [23] and Grigg et al. [63], whilst mounds featuring one to two openings correspond to observations by Stapleton et al. [25] and Grigg et al. [63].
In addition to modifying bottom topography, sediment reworking and burrow creation by T. australiensis also induced differences in surface sediment profiles, with the sediment character of 0-4 cm sediments typically differing from the deeper sediments ( Figure 8). Differences between the upper and deeper sediment characters reflect the influence of the increased sediment reworking (burrow maintenance and excavation) within the upper sediments.

Sediment Reworking by T. australiensis
Luminophores and a gallery-diffusion model were used to provide quantitative estimates of sediment reworking. Luminophore tracers remained on the sediment surface in the non-bioturbated mesocosms (Figure 3a,d), whilst in the bioturbated mesocosms, the majority of the luminophores were buried in the top 5 cm, but otherwise were distributed quite evenly down to 55 cm (Figure 4). The biodiffusion like/non-local resulting combination pattern validated the use of the gallery-diffusion model to quantify sediment reworking [38,51,64] (Supplementary File 4). Surface biodiffusion-like sediment reworking by T. australiensis was higher than reported for many other intertidal benthic organisms (including polychaetes, echinoderms and bivalves). Specific examples of biodiffusion coefficients include 1.1 to 2.0 cm 2 y −1 for ragworms (Hediste diversicolor) [65], 0.59 cm 2 y −1 for T headed worms (Scalibregma inflatum) [51], 2.2 cm 2 y −1 for the sea potato (Echinoidea cordatum) [51], 0.17 to 51.63 cm 2 y −1 for the bivalve Abra ovata [66] and 0.82 cm 2 y −1 for Baltic clams (Macoma baltica) [67]. Additionally, Cooper [68] reported mean biodiffusion coefficient estimates for the sandprawn Callichirus kraussi, a species of ghost shrimp and conveyor belt feeder with very similar anatomy to T. australiensis, to range between 0.57 and 1.59 cm 2 y −1 . During this study, the presence of mangrove leaf litter did not significantly influence sediment reworking activity by T. australienis (Table 1). Apparent biodiffusion is due to the movement and feeding activity of organisms, the creation and maintenance of burrows, and passive transport due to tracers falling into burrow openings [38,54,68]. Non-local transport below the biodiffusive layer is the result of particle translocation within burrow structures, such as egestion of faecal pellets and burrow maintenance [28]. Such deep sub-surface transport was also generated by T. australiensis down to 55 cm depth with an intensity comparable to those previously reported for C. kraussi (up to 3.5 y −1 [67]).
T. australiensis actively search for the particulate organic matter within sub-surface sediment layers, mixing particles as they move within and extend their burrows [25,69]. T. australiensis burrows typically consist of one or more surficial U-shaped sections in the upper 10 cm of sediment, which link the burrow openings with a sub-surface "turning" chamber. A deep, vertical shaft with lateral chambers or tunnels descends from this turning chamber to a depth of 40 cm or more [25,70]. Therefore, T. australiensis would be expected to redistribute tracer particles to depth and theoretically, a high density of U-shaped burrow sections in the upper 10 cm of sediment would increase biodiffusive mixing, as D b has been shown to have a density-dependent relationship with mixing [64,71,72].
Given their relatively large size, reworking rates, and high burrow densities, which may be as great as 500-1000 m −2 [33], T. australiensis is expected to play a significant role in engineering physical and biogeochemical conditions in its local environment.

Physical Sediment Characteristics
T. australiensis significantly influenced sediment bulk density with depth-averaged densities in the bioturbated treatments lower than the non-bioturbated sediments. Additionally, bioturbated treatments demonstrated higher sediment porosity than those in the non-bioturbated treatments ( Table 2, Supplementary File 1). Such findings agree with previous studies regarding burrowing shrimp [8,73] due to reworking and turnover of sediments altering the size of the interstitial spaces.
Continuous sediment reworking (burrow maintenance and feeding behaviours) that resulted in the dynamic nature of the mound placement may also act as a strategy to capture and enhance organic matter supply to the sub-surface sediments, providing additional food resources that would otherwise remain on the sediment surface. This would prove particularly beneficial where mangroves as leaf litter that fall onto the sediment surface would be trapped and buried below the burrow mounds in yabby inhabited environments.
Burrows create an uneven sediment topography can also alter bottom current regimes [74] and influence critical shear stress for erosion, particle size distributions and sediment resuspension [31,[75][76][77] and can increase the settling of organic particles [78], again potentially supplementing the diet of the bioturbating fauna.

Sediment Organic Matter and Nutrient Profiles
Following the initial 17-day mesocosm stabilisation period ( Figure 2) the physical sediment characteristics and nutrient concentrations in the sediments resembled those of natural sandy sediments of the Gold Coast Broadwater [5,[79][80][81][82][83]. As such, whilst diurnal or seasonal variations, and competing infauna interactions were not considered, interpretations of our results can be considered appropriate to natural conditions. The activities of T. australiensis appear to ensure that organic matter does not simply accumulate on the sediment surface, but rather is translocated to depth, or simply buried due to the expulsion of sediment from the burrows (Section 3.1). This indicates the potential for T. australiensis to actively increase organic matter accumulation in intertidal coastal sediment, especially near mangrove forests where leaves are abundant (at low tide falling leaves are deposited directly onto the sediment surface or are stranded there by the ebbing tide). This behaviour would be particularly influential in environments, where organic detritus would otherwise remain mobile on the sediment surface.
Ghost shrimp species have also been shown to actively collect large particulate organic matter from the sediment surface and transport it into their burrows [10], influencing benthic metabolism, bacterial abundance and nutrient recycling. However, in this study, despite the rapid burial of introduced leaf litter and the obvious presence of buried leaf litter in sectioned core samples at the end of the incubations, there was no significant difference in sediment organic matter (% organic C and LOI 550 ) content between the mesocosm treatments. This may reflect the fact that benthic fauna, including T. australiensis, also enhance benthic respiration and organic matter turnover in sediments [14,23,30,32]. Consequently, the net effect of T. australiensis on sediment organic matter pools will depend on the relative degree to which organic matter inputs are stimulated compared to the degree to which organic matter losses via microbial respiration are stimulated. The net balance between these processes is also likely to be strongly influenced by local factors, such as the amount and types of organic matter present, yabby population density and the sediment type [10,23,84]. Thus, the lack of effect of T. australiensis on sediment organic matter pools may indicate that organic matter inputs through burial of leaf litter were largely offset by increased losses of sediment organic matter via respiration. This conclusion is supported by the shift recorded in the C:N ratio of the sediment organic matter pools in this study. At the end of the mesocosm incubations (day 55) C:N ratios were significantly different between treatments, with C:N ratio profiles within the bioturbated mesocosms demonstrating higher ratios in the upper 2.5-20 cm depth horizon of the S+Y+OM treatment in comparison to all other treatments. This shift in the C:N ratio towards that of the added mangrove leaf litter of 36.9 ± 2.4 [32] indicates that the added leaf litter was accumulating within these bioturbated sediments even if this was not reflected by significantly greater % organic C or LOI 550 in the sediment. This also suggests that the field site received organic inputs from a mix of sources.
Measured δ 13 C values were similar to those reported within the Gold Coast Broadwater system [5]. Mean δ 13 C values significantly differed between depths (Table 2), with values increasing with depth. Within the 0-20 cm depth, S+Y+OM δ 13 C values were predominantly depleted in comparison to other treatments potentially reflecting the influence of mangrove leaf litter signature [5] translocated by T. australiensis sediment reworking. Additionally, δ 15 N values measured during this study reflect those reported for sandy sediment within the Gold Coast Broadwater system [5]. Similarly, δ 15 N values significantly differed between depths ( Table 2) with decreased values in the deeper sediment. Although no clear trend relating to the influence of T. australienesis was apparent, δ 15 N values in the bioturbated sediments (S+Y and S+Y+OM) demonstrated a greater change relative to the initial δ 15 N values (conditions) in comparison to the non-bioturbated sediments in all but the deepest depth layer. Shifts in values are potentially attributable to a variety of possible fractionation processes, including bacterial colonisation, and processes related to ammonification, nitrification, denitrification and nutrient assimilation reactions, which have been shown to be influenced by T. australiensis [30,32].
Numerous studies have shown that benthic infauna profoundly influences nutrient cycling processes and dissolved nutrient fluxes between the sediment and water column due to their bioturbation and sediment bioventilation [14,[85][86][87][88]. In this study T. australiensis significantly influenced sediment NH 4 + bio concentrations with decreased concentrations of NH 4 + bio recorded in the bioturbated S+Y and S+Y+OM treatments compared to the non-bioturbated S and S+OM treatments. In non-bioturbated sediments, NH 4 + bio concentrations typically increased with increasing depth, whilst bioturbated sediments exhibited lower and more consistent NH 4 + bio concentrations over the entire depth profile (Figure 7). Similar decreases in sediment NH 4 + concentrations have been observed in association with burrowing macrofauna [55,59,89], including T. australiensis [23,30,32]. These decreases in ammonium concentrations in bioturbated sediments can be directly related to the bioventilation activity of sediment infauna, which introduces oxygenated water into the deeper sediment strata, favouring oxidation of ammonium to nitrite and nitrate via nitrification, and also flushes porewater ammonium to the overlying water [14,87]. Increased rates of nitrification and ammonium efflux to the water column in the presence of T. australiensis have previously been shown to produce lower sediment NH 4 + bio concentrations during mesocosm manipulations even though T. australiensis also stimulated ammonium production (ammonification) rates in the sediment [30,32]. Therefore, it is highly likely that the lower sediment NH 4 + bio concentrations recorded in the bioturbated mesocosms in this study reflect similar changes in the net balance between nitrification, ammonium efflux and ammonification rates.

Conclusions
This study has demonstrated that T. australiensis populations have a significant influence on the surface and sub-surface sediments. Furthermore, the findings corroborate that T. australiensis has an important role as an 'ecosystem engineer' in intertidal habitats through the burial and 'trapping' of organic detritus through vigorous and dynamic burrowing activity. While this study used mangrove leaves, we anticipate this behaviour would be effective with many forms of organic detritus. Observed influences of T. australiensis on physical and chemical sediment conditions, included changes to topography, the rapid burial of introduced leaf litter influencing the distribution and organic matter availability, whilst also significantly influencing sediment bulk density and porosity. Increased porosity plays a central role in sustaining high permeability of surface sediments, promoting porewater exchange, influencing organic matter remineralisation and nutrient cycling. Furthermore, T. australiensis significantly altered NH 4 + bio concentrations, influencing nutrient cycling and sediment-water interface fluxes. It is suggested that efforts be made to effectively manage T. australiensis populations from pressures such as commercial and recreational (bait) harvesting as the loss of T. australiensis in (previously) inhabited sediments could potentially have cascading effects on ecosystem conditions. The over-exploitation of such ecosystem engineers presumably would have consequences that extend beyond their own decline, to also influence sediment characteristics and biogeochemical functions.