Hardened faecal pellets as a significant component in deep water, subtropical marine environments

Non‐skeletal carbonate grains are classically interpreted to form in shallow, tropical environments. Peloids deposited in deep, subtropical marine conditions are poorly studied. IODP site U1460 on the subtropical Carnarvon Ramp (Southwest Shelf of Australia) recovered a nearly continuous Pliocene to Recent record of outer shelf and slope sediments. The relative abundance of peloids varies between 0% and 67% of the fine to medium sand fraction, and contributes on average ~4% of all grains. The origin and composition of these peloids were investigated using scanning electron microscopy equipped with an energy‐dispersive X‐ray spectrometer, light microscopy, X‐ray diffraction and stable isotope analysis. The peloids have a uniform size and shape and are interpreted as faecal pellets. They are mainly composed of skeletal fragments such as ascidian spicules, planktic foraminifera and sponge spicules in a mud‐sized matrix containing abundant coccolith plates. Mineralogical analysis show that the pellets consist of aragonite, calcite and dolomite. The pellets have an identical mineralogical composition and skeletal assemblage as the surrounding matrix, indicating that they have formed in situ. They occur more abundantly during interglacials when the site was situated in deeper waters below the swell wave base, presumably because the pellets were protected from disintegration and therefore available for cementation. The presence of framboidal pyrite within the pellets indicates bacterial sulphate reduction (BSR). The reduction of iron by hydrogen sulphide produced during BSR decreases the pH and likely explains the observed aragonite dissolution. Aragonite dissolution likely increases the alkalinity, and in consequence causes the precipitation of calcite and dolomite cements. It is suggested here that pellets are hardened due to this early cementation close to the sea floor increasing the potential for preservation in the fossil record.


| GEOLOGICAL SETTINGS
The environment of the south-western continental margin of Australia in the eastern Indian Ocean is transitional between warm-temperate and tropical carbonate settings. It comprises the warm-temperate Rottnest Shelf in the south (south of 28°S) and the subtropical Carnarvon Ramp in the north (22-28°S). The Houtman Abrolhes Reef complex, which contains the southernmost major tropical reef in the Indian Ocean (28-29.5°S), straddles the boundary between the Carnarvon Ramp and the Rottnest Shelf (James, Collins, Bone, & Hallock, 1999; Figure 1). The modern-shelf sediments are characterized by a distinct cool-water composition (Nelson, 1988), but F I G U R E 1 (a) Location map of IODP site U1460, Tamala line (N, red circles) and sedimentary facies on the modern sea floor of the Carnarvon Ramp/northern Rottnest Shelf, along Southwest Shelf of Australia. Bathymetric contours are in metres (modified after James et al.,1999). (b) Cross-section of Tamala line (N) with distribution of modern facies and location of the IODP site U1460 (modified after James et al., 1999) with subtropical attributes (James, Bone, Hageman, Feary, & Gostin, 1997). The skeletal assemblage is dominated by coralline algae, bryozoans, molluscs (scaphopods, bivalves and gastropods) and foraminifers (James et al., 1997). The main difference from typical cool-water deposits is the presence of scattered zooxanthellate corals and large, symbiontbearing foraminifers (Collins, James, & Bone, 2014;James et al., 1999). Skeletons of serpulid worms, echinoids, azooxanthellate corals and sponge spicules are of local importance. The calcareous green algae Halimeda grows locally on the shelf and ramp but is poorly calcified and therefore generally does not contribute to the sediment outside Shark Bay. The sea floor above ~50-60 m is subject to constant reworking and abrasion by waves, while the swell wave base is close to 100 m, leading to a lack of mud deposition above this depth.
The Carnarvon Ramp comprises the Ningaloo Reef and hypersaline Shark Bay on the inner ramp. The mid-ramp is euphotic, with relatively little calcareous benthos and with low numbers of bryozoans, coralline algae and larger foraminifers, but is dominated by relict or stranded foraminiferaldominated sand. The outer ramp is pelagic in character, covered by planktic foraminiferal sand or spiculitic mud. The depression of temperature and salinity contours indicates the influence of periodic downwelling and the outflow of highly saline waters from Shark Bay (Collins et al., 2014;James et al., 1999). The Rottnest Shelf is flat-topped, with a waveswept inner shelf plain characterized by rhodolith pavements, bryozoans, sponges and abraded sediments. An incipient rim formed by a linear ridge system covered by rhodolite gravel separates the inner shelf from the subphotic outer shelf that is dominated by bryozoans, benthic foraminifers and molluscs. The upper slope contains fine sand and silt of bryozoan fragments, sponge spicules and planktic foraminifers (James et al., 1999).
IODP site U1460 was drilled during IODP Expedition 356 on the Southwest Shelf of Australia in a water depth of 214.4 m at 27°22.4867′S and 112°55.4265′E (Figure 1). It is situated north of the Houtman Abrolhos reef complex, at the transition between the Carnarvon Ramp in the north and the Rottnest Shelf towards the south (Figure 1a,b). This transition zone shows a mixture of morphologic and facies characteristics of both regions. James et al. (1999) described the shelf morphology and facies distribution on the modern sea floor. The facies on the inner shelf is characterized by skeletal sands (facies A1; Figure 1) and gravel, with the siliciclastic content increasing towards the coast (facies A2; Figure 1). A high abundance of intraclasts and other relict and stranded grains characterizes the nearshore zone down to a water depth of ~50 m (facies I1; Figure 1). The ridge complex in the transition zone is deeper and morphologically less well defined compared to the Northern Rottnest Shelf. It separates the relatively deep inner shelf from a more gently inclined outer ramp/shelf. The ridge complex, which is interpreted as being a late Pleistocene beach-dune complex (James et al., 1999), is covered by rhodolite gravel and subordinate skeletal sand (facies A3; Figure 1). In the transition zone, the sediment on the ridge complex is relatively rich in bryozoans. The outer ramp/shelf extends from the base of the ridge complex to the shelf edge at approximately 200 m water depth (mwd). The dominant facies on the shallower part of the outer ramp/ shelf is bryozoan skeletal sand (facies B; Figure 1), while the deeper part is characterized by pelagic sand facies (P1, P2). The pelagic sand facies is situated below the swell wave base and therefore is the shallowest facies containing carbonate mud (>20%). The content of azooxanthellate corals in this facies is relatively high, with a ~10-20% contribution to the coarse fraction (James et al., 1999). The continental slope in more than 200 mwd is dominated by the carbonate silt facies (S), containing ~70% mud, 30% sand and only trace amounts of coarse fraction (>2 mm). The sand fraction is composed mainly of 95% carbonate with trace terrigenous grains. Planktic foraminifers and sponge spicules contribute most skeletal elements to the sand fraction, while bryozoans, ostracods, pteropods and benthic foraminifers are also common. Worm tubes and faecal pellets are further sediment constituents that can be locally abundant. Azooxanthellate corals, in contrast, seem to be absent (James et al., 1999).

| MATERIAL AND METHODS
From Hole U1460B, a 306 m long hydraulic piston core with 98% recovery was obtained ( Figure 2). The IODP site U1460 consists of two lithostratigraphic units (Ι, ΙΙ), with Unit I subdivided into three subunits (Ιa, Ιb, Ιc) (Gallagher et al., 2017). In this study, the focus is placed on the interval between 0 and 108 m core depth below sea floor (CSF-A) belonging to Subunits Ia (0-44.94 m CSF-A) and .36 m CSF-A) ( Figure 2).
Unit I as defined in Gallagher et al. (2017) consists mainly of unlithified to partially lithified skeletal packstone with minor skeletal wackestone, and some lithified intervals ( Figure 2). This unit is divided into three Subunits based on the abundances of macrofossils and sponge spicules and variation in diagenetic alteration. Macrofossils are concentrated in Subunit Ia, and sponge spicules in Subunit Ib (Gallagher et al., 2017). Subunit Ia is characterized by beige to greenishgrey skeletal packstones. The packstones are interbedded with skeletal wackstones and mudstones in the uppermost 20 m (CSF-A). The sediment is predominantly unlithified, except for a minor lithified interval in the lowermost part of Subunit Ia. Bioclasts of neritic macrofossils such as echinoderms, bryozoans, bivalves and gastropods are present, with major pelagic components including planktic foraminifers. The base of the lithified interval in Subunit Ia is defined by a hardground. The upper part of Subunit Ib is dominated by | 351 DEIK Et al.
unlithified beige to light brown skeletal packstone and the lower part by partially lithified skeletal wackestone and mudstone (Gallagher et al., 2017).
The cores U1460-1F to 25F (0-108.6 m CSF-A) were sampled with a 10 cm 3 tube every 20 cm on average. This interval represents a time from 0 to 762 ka, following the age model of Petrick et al. (2018), who tuned biomarker derived Sea Surface Temperature estimates to the benthic isotope stack LR04 at this site. This age model is also consistent with the last occurrence of the planktic foraminifer Globorotalia tosaensis (0.62 Ma) at 86.5 m CSF-A (Gallagher et al., 2017). Marine isotope stages (MIS) were assigned according to Petrick et al. (2018).
All 103 samples were wet sieved through a 63 μm sieve, and the retained coarser fraction was subsequently dry sieved through 125, 500 and 2,000 μm sieves. The mud fraction (<63 μm) of a subset of 24 samples was further wet sieved through a 34 μm sieve. All individual grain-size fractions of these samples were weighed. A small portion of the bulk sediment from the same subset was obtained for mineralogical analysis using X-ray diffraction (XRD). Samples were oven-dried, ground and mounted on sample holders. The measurements were conducted using a Siemens D5000 X-ray diffractometer over an angle field of 60° (4°-64°) with a step size of 0.02° per second. Identification and quantification of different mineral phases was achieved with the software DIFFRAC EVA (ver. 8.0) by Bruker. The relative abundance of mineral phases was determined using the I/I corundum values from the International Center for Diffraction Data database. Non-destructive determination of mineralogy of peloids was achieved by 2-D XRD measurements (Smodej et al., 2015).
The abundance of peloids in the 125-500 μm fraction of all 103 samples was quantified with a point counter by F I G U R E 2 (a) Lithostratigraphic summary for IODP Hole U1460B. Lithification varies between unlithified (1), partially lithified (2) and lithified (3) intervals (Gallagher et al., 2017). The presence of azooxanthellate corals is displayed in (b). Carbonate (blue) versus siliciclastic (black) content (c). (d) Grain size fraction distribution shows a downward increase of the mud fraction (grey). Sand fraction (yellow), gravel (brown). Carbonate minerals are displayed in e (aragonite: black, low-Mg calcite: grey, dolomite: dark grey and the high-Mg calcite: light grey). XRD results (f) show a systematic decrease in aragonite content in bulk sediment with depth. (g) Relative abundance of peloids in the fine to medium sand-sized (125-500 μm) sediment fraction. (h) MIS stratigraphy (Petrick et al., 2018) and biostratigraphic age marker (i). Grey bars indicate the hardground intervals. (j) δ 13 C (red square) and δ 18 O (blue dot) values of peloids from IODP site U1460 in ‰ V-PDB counting 300 grains per sample. The length and width of 89 peloids from a total of nine samples was measured under the binocular microscope. A total of 16 thin sections from the first 70 m (CSF-A) were prepared for petrographic analysis of the sand and gravel fraction.
Scanning electron microscopy was used to analyse the composition of pellets and the grain fractions <34 μm and 34-63 μm. The mineralogy of grains was confirmed by elemental analysis (Sr, S, Fe, Mg) with an energy-dispersive X-ray spectrometer (EDS). All samples were carbon coated prior to analysis.
Ten peloid samples from the upper 40 m (CSF-A) were selected for stable isotopic analyses (δ 18 O and δ 13 C), which were performed at Leipzig University. Carbonate powders were reacted with 105% phosphoric acid at 70°C using a Kiel IV online carbonate preparation line connected to a MAT 253 mass spectrometer. Isotope values were calibrated to the Vienna Pee Dee Belemnite (V-PDB) standard using the NBS-19 carbonate standard. Average standard deviation associated with the analyses of a reference standard is <0.050‰ V-PDB or δ 18 O and <0.016‰ V-PDB for δ 13 C.

| Composition, mineralogy and texture of sediment
The average carbonate content of the sediment is ~90%, with 57% low-magnesium calcite (LMC), 12% aragonite, 12% high-magnesium calcite (HMC) and 9% dolomite (Table 1; Figure 2e). The aragonite content of the bulk sediment decreases with depth from around 30% to 4% at the base of the section ( Figure 2f). Above ~50 m (CSF-A), the mean HMC content is 26%, but HMC disappears completely below this depth. At around the same depth, the dolomite content increases slightly and reaches a maximum value of 21% at a depth of 94.5 m CSF-A in the lower half of the studied interval. The siliciclastic fraction consists mainly of quartz and plagioclase feldspar (Figure 2c). Sulphate minerals such as celestite and anhydrite occur locally as a minor component.
Subunits Ia and Ib were defined onboard the Joides Resolution based on differences in skeletal assemblage, mineralogy and diagenetic alteration (Gallagher et al., 2017). This subdivision is confirmed by the data presented here. Subunit Ia is characterized by a relatively high content of aragonite (19% , Table 1) and HMC (25%, Table 1). Based on visual estimates, the skeletal assemblage is dominated by planktic foraminifers, mollusc fragments (including pteropods), bryozoans, echinoids and some azooxanthellate corals, intraclasts and peloids. The sediment is a mixture of sand (64% , Table 1) and mud (33%, Table 1). The gravel content is very low (3% ,  Table 1), except for some coarse-grained intercalations that consist of sand (63%) and gravel (37%). The sand fraction in these layers consists predominantly of bryozoan fragments, the rest comprising variable proportions of foraminifers, molluscs, serpulids, echinoids and minor amounts of azooxanthellate corals, ascidians spicules and scaphopods. The gravel fraction has a similar composition but contains many skeletal intraclasts and is lacking ascidian spicules. Peloids and quartz grains are generally absent in all of these coarsegrained intercalations.
The investigated part of Subunit Ib is characterized by a lower aragonite (~7%, Table 1) content compared to Subunit Ia and by a near absence of HMC as well as an increased dolomite content (~12%, Table 1). The skeletal assemblage is similar to Subunit Ia, but the number of macrofossil fragments is generally lower. Further differences are the high content of sponge spicules, which were much less in subunit Ia, and the absence of azooxanthellate corals ( Figure 2). The content of peloids at Subunit Ib increases compared to Subunit Ia (14%,

| Morphology and composition of peloids
The ovoid to ellipsoid shaped peloids are generally medium to coarse sand-sized, with a mean length of ~650 μm (SD: 90 μm), a mean width of ~390 μm (SD: 60 μm) and a length to width ratio of ~1.7. Because of their width, the peloids are found nearly exclusively in the 125-500 μm fraction. Over the entire interval, peloids contribute a mean of ~10% grains (range: from 0% to 67%; Figure 2g) to this size fraction (Table 1), or 4% of all grains. Their content nearly doubles from Subunit Ia (~7%, Table 1) to Subunit Ib (~14% ,  Table 1). Peloid abundance is always higher in each interglacial stage (odd MIS) compared to the overlying glacial stage (even MIS; Figure 2, Table 2). This relationship is most clearly visible in the upper ~30 m CSF-A.
The peloids have a smooth surface, very uniform morphology and size ( Figure 3) which characterize them as faecal pellets (Blom & Alsop, 1988;O'Connell & James, 2015). They are mainly composed of skeletal fragments such as ascidian spicules, planktic and benthic foraminiferal tests, and sponge spicules in a micritic matrix containing coccolith plates (Figure 4b-f). Non-carbonate minerals such as pyrite and plagioclase feldspar occur as further locally important constituents. 2D-XRD shows that the pellets have a very similar mineralogical composition to the bulk sediment at the same depth.

| Diagenetic processes and stable isotopes
Aragonite dissolution in the mud-sized matrix and within the faecal pellets is visible at the tips of ascidian spicules (Figures 4c and 5a). Authigenic phases formed within the matrix sediment and the faecal pellets are framboidal pyrite, HMC (7-9 mol%) and dolomite cement (Figures 5b-d and  6a-e). Phosphate was observed lining the outer rim of faecal pellets (Figure 6f).

| Origin of peloids
At present, IODP site U1460 is situated in a water depth of ~214 m, at the transition between the planktic sand and silt facies (P2) and the carbonate silt facies (S1). A comparison with the composition of present sea floor sediments (James et al., 1999) indicates that the sediments in Subunit Ia are very similar to the planktic sand and silt facies (P2; James et al., 1999), i.e. consisting of abundant planktic foraminifers, relatively few intraclasts and ~33% mud content. On the modern day sea floor, this facies is restricted to the outer ramp and occurs in water depths of between ~120 and 200 m (Figure 1). The majority of the sediment in subunit Ib is therefore interpreted to have formed over a similar depth range. The coarse-grained, gravel-rich sediments, in contrast, are similar in composition to the bryozoan skeletal sand facies (B; James et al., 1999) on the modern day sea floor. The main similarities to this facies are the amount of gravel, the lack of mud, the absence of quartz grains and the high proportion of bryozoan debris compared to coralline algae fragments. The bryozoan skeletal sand facies is typically deposited on the outer shelf and occurs in water depths of between 100 and 150 m, landward of Site 1460 (Figure 1, James et al., 1999). Based on their correlation with glacial MIS (Figure 8), at least some of the gravel-rich layers were likely deposited in situ during glacial sea-level lowstands and therefore do not represent drilling disturbances (fall-in deposits) as interpreted by Gallagher et al. (2017) in the shipboard reports.
The investigated part of Subunit Ib shows many similarities with the spiculitic carbonate silt facies (S1) such as the high mud content, the abundance of sponge spicules and the absence of azooxanthellate corals. The spiculitic carbonate silt facies is also the only facies from which peloids are described on the modern sea floor (James et al., 1999). Today, this facies occurs in water >200 m deep on the slope of the Carnarvon ramp and the northern Rottnest Shelf (Figure 1). Therefore, Subunit Ib was likely deposited in deeper water compared to Subunit Ia. This is also in accordance with the lower aragonite content in Subunit Ib (Figure 2), as there is a tendency for decreasing aragonite content in modern coolwater environments from the middle and outer shelf towards the upper slope (James, Bone, & Kyser, 2005). Additionally, the decrease in aragonite with depth could be caused by subsea floor aragonite dissolution . Sediments of Subunit I were deposited during several glacial-interglacial cycles, which are expected to influence the skeletal assemblage in the inner and mid-ramp/shelf environments. However, the climate-related changes in skeletal assemblages in the outer ramp/shelf and slope sediments deposited at IODP site U1460 are expected to be subtle since bottom water temperatures at the site were likely always <15°C (James et al., 1999).
The similarity of the skeletal assemblage in both the matrix and the pellets indicates that pellets have formed in situ and were not imported from shallower environments. The abundance of skeletal elements from benthic organisms such as sponge and ascidian spicules indicates that they were formed by deposit-feeders and not by pelagic organisms. The fact that all pellets are very similar in composition, size and shape point to a single group of producers. The ovoid to ellipsoidal shape with a length to width ratio of 1.7:1 is typical for faecal pellets of polychaete worms or molluscs (Bandel, 1974;Martens, 1978). Deposit feeding polychaete worms occur abundantly on the Carnarvon ramp (Brooke et al., 2009) and even form the most prominent group of benthic organisms on soft, fine grained sediments on the Northwest Shelf of Australia (Jones et al., 2007). It therefore seems to be likely that polychaete worms like are the producers of the faecal pellets. Their higher abundance of pellets in interglacial intervals indicates that more pellets were deposited when IODP site U1460 was situated in deeper waters. Additionally, the presented facies interpretation indicates that pellet abundance is higher in the deeper water slope facies (Subunit Ib, 14% of 125-500 μm fraction) than in the more proximal outer ramp/shelf facies (Subunit Ia, 7% of 125-500 μm fraction).
T A B L E 2 Mean peloid abundance in the fine to medium sand fraction (125-500 μm) for each Marine Isotope Stage (MIS). N is the number of samples within each MIS and SD is the standard deviation

| Aragonite dissolution
Bioclastic aragonite at IODP site U1460 was produced mainly from the shells and skeletal elements of ascidians, pteropods and bivalve shells via maceration. Maceration is the breakdown of grains along lines of weakness between skeletal subunits into their microscopic structural elements, needles and granules (Figure 6c; Alexandersson, 1979;O'Connell & James, 2015). The observed systematic decease in aragonite content with depth could be due to platform progradation (see above) and dissolution during shallow burial (Figure 2). The presence of the framboidal pyrite in pellets already at their shallowest occurrence (1.15 m CSF-A) indicates active bacterial sulphate reduction (BSR) near the sea floor.

| Stable isotope signature of peloids
The studied faecal pellets from the outer shelf of the subtropical Carnarvon ramp show higher oxygen and lower carbon isotopes values compared to their tropical counterparts, e.g. from the modern arid carbonate ramp of Kuwait or the isolated carbonate platforms of the Belize-Yucatan system (Gischler & Lomando, 2005;Gischler, Swart, & Lomando, 2009) (Figure 7). In contrast, the values are very similar to isotope values of sea floor carbonate sediments and HMC cements from the cool-water, southern Australian Shelf (Figure 7). The isotopic signature therefore might be a useful discriminator for pellets from these two environments. 6 | DISCUSSION

| Origin of peloids
Non-skeletal grains are typically formed in shallow, tropical environments such as the Bahamas (Bathurst, 1975;Harris, 1979;Hine, 1977;Lloyd, Perkins, & Kerr, 1987;Purdy, 1963). Sediment sampling at the sea floor indicated the presence of faecal pellets on the slope of the Southwest Shelf of Australia at a depth >200-300 m (James et al., 1999). Several previous studies on temperate to cool-water regions indicate that faecal pellets disintegrate easily into mud-sized carbonates (Farrow & Fyfe, 1988) and that hardened faecal pellets occur only sporadically in temperate, shallow marine carbonates (O'Connell & James, 2015). This is consistent with the lack of faecal pellets in waters shallower than 200 m on the swell-dominated Southwest Shelf of Australia (James et al., 1999). However, Blom and Alsop (1988) have demonstrated that faecal pellets accumulate in relatively shallow water at depths of between 70 and 85 m in the central Bass Basin (Southeast Australian Shelf), where they are protected from swells. The data presented here show that faecal pellets form a significant part of the sand fraction of the middle to upper Pleistocene interval at IODP site U1460. The faecal pellet abundance shows a clear covariance with relative water depth since pellets occur preferentially in the slope facies (Subunit Ib) and during interglacial intervals when IODP site U1460 was situated in relatively deep water. The changes between glacial and interglacial intervals are more clearly expressed in the upper ~30 m CSF-A (MIS 1-12), which was influenced by high-amplitude sea-level variations. For example, during the Last Glacial Maximum, when sea-level dropped by ~120 m (Lisiecki & Raymo, 2005), the water depth at IODP site U1460 would have been around ~90 m. This would have brought the site above the present day swell wave base (James et al., 1999), causing pellets to disintegrate (Figure 8b; glacial MIS1-12). Compared to previous glacial-interglacial cycles, MIS 13-15 were possibly characterized by relatively lower amplitude sea-level fluctuations (Miller et al., 2005), insufficient to place the site above the swell wave base and potentially explaining the lack of a clear correlation between MIS and pellet abundance in this interval. A more important reason might be that the depositional environment was generally deeper in the lower part of the interval, as indicated by a shift to slope facies in Subunit Ib, dampening the effect of sea-level fluctuations.

| Comparison to peloids from tropical environments
This study shows that faecal pellets can be formed and preserved in temperate carbonates on the transition from the outer ramp/shelf to slope, where they are protected below the swell wave base from disintegration. Volumetrically, they form up to 67% of the fine to medium sand fraction and on average ~2% of the total sediment of the entire interval. The abundance of pellets in these relatively deep-water temperate carbonates is therefore similar or higher than most tropical carbonate shoals and shallow water platforms in the Indo-Pacific region (Gischler, 2011;Utami, Reuning, & Cayharinin, 2018). Many carbonate platforms and shoals in the Caribbean (Milliman, 1967(Milliman, , 1969 have higher peloid contents compared to IODP site U1460, but several have similar or even lower peloid abundances, for example the Dry Tortugas shoal on the distally steepend ramp of the west Florida shelf or the Turneffe Islands platform off Belize F I G U R E 8 Bathymetric profile across the outer ramp/shelf (modified after James et al., 1999, for position see Figure 1). Facies distribution along the profile (compare Figure 1) is indicated for carbonate silt (S), pelagic sand+silt (P) and bryozoan skeletal sand (B) facies. IODP site U1460 is located below the modern swell wave base (A), but would be located close to the swell wave base during the Last Glacial Maximum (B). Facies boundaries and the swell wave base in B are shifted downward by ~120 m along with the sea-level fall, assuming no major climate-related facies change at the relatively deep water in the vicinity of the upper continental slope F I G U R E 7 δ13C and δ18O values of faecal pellets from IODP site U1460 and of peloids from Kuwait ramp and Belize-Yucatan platforms (Gischler & Lomando, 2005;Gischler et al., 2009). Isotopic composition of sea floor skeletal sediments (Gillespie & Nelson, 1997) and HMC cements (Rahimpour-Bonab, Bone, Moussavi-Harami, & Turnbull, 1997) from cool-water environments are shown for comparison (Gischler, Isaack, & Hudson, 2017;Gischler & Zingeler, 2002). The skeletal assemblage contained in those shallow water, tropical faecal pellets is partially similar to the pellets from deeper water, temperate carbonates described in this study. Gischler (2011) mentioned that cemented faecal pellets from Bora Bora contain skeletal components such as mollusc and (benthic) foraminiferal shell fragments, sponge needles or tunicate spicules. However, the tropical shallow water pellets lack the pelagic component that is common in the pellets at IODP site U1460. Another difference beside the skeletal pelagic component is the mineralogy of the pellets. Similar to the bulk sediment, the mineralogy in tropical pellets is dominated by aragonite versus calcite at IODP site U1460. Tropical peloids are cemented by aragonite, while pellets at IODP site U1460 show HMC and dolomite cements.
A clear difference is also visible in the isotopic signature of the faecal pellets. The different carbon isotope values can partly be explained by mineralogical differences. Rubinson and Clayton (1969) have shown that the enrichment of δ 13 C in aragonite is about 1.8‰ higher compared to LMC. The aragonite content is about 80% higher at the tropical sites (Gischler, 2006;Gischler & Lomando, 2005) compared to IODP site U1460. The different aragonite content therefore could account for ~1.44‰ of the lower δ 13 C values seen in the pellets at IODP site U1460. However, the observed difference is on the order of ~3‰, indicating that aragonite concentration alone cannot explain the lower δ 13 C. Therefore, the higher δ 13 C values in tropical peloids might at least partially reflect the isotopic enrichment of the dissolved inorganic carbon pool due to photosynthesis observed in environments with a low water exchange rate (Swart & Eberli, 2005;). Carbonate cement precipitated under the influence of BSR, assumed to occur with the studied pellets, can show δ 13 C depletion (Machel, 2001). However, this effect is typically restricted by the buffering effect of the carbonate matrix and therefore is less important in this case. The most likely reason for the difference in δ 13 C is the higher content of skeletal material in pellets at IODP site U1460 that typically form in isotopic disequilibrium and generally have lower δ 13 C values (Gischler et al., 2009;Swart et al., 2009). The discrepancy in the δ 18 O signature between the pellets from the tropical sites and IODP site U1460 can be explained by difference in salinity and temperature. Temperatures and salinities are more elevated on the carbonate ramp of Kuwait (39‰-43‰, 13-32°C; Gischler & Lomando, 2005) and the carbonate platforms of the Belize-Yucatan system (35‰-42‰, 27-28°C; Gischler, Hauser, Heinrich, & Scheitel, 2003) compared to the environmental conditions on the outer ramp and slope of the Southwest Shelf of Australia (35. 6‰-35.8‰, 17-18°C;James et al., 1999). Assuming a δ 18 O-temperature relationship of 0.21 ‰/°C (Bemis, Spero, Bijma, & Lea, 1998), the different δ 18 O values between IODP site U1460 and the Belize-Yucatan Platforms (~2.2 ‰) would indicate ~11°C lower temperatures at IODP site U1460. The match between these calculated and observed temperature differences indicates that much of the bioclastic sand and mud incorporated in the pellets was formed and cemented in relatively cool-bottom waters. This is also consistent with the similarity to the isotopic signature of HMC cements formed on the sea floor of the cool-water Lacepede Shelf (Figure 7). Higher δ 18 O water values in the arid, hypersaline setting of the Kuwait ramp likely explain the lower δ 18 O difference between pellets from that site and IODP site U1460. Evaporation leads to the preferential removal of 16 O from surface waters and hence to an increase in δ 18 O. This is often termed the 'salinity effect', since the salinity is affected by the same processes as the isotope fractionation in sea water. The observed offset in carbon and oxygen isotope values between IODP site U1460 and the modern examples is stable over the investigated interval (upper ~40 m; ~500 kyr; Figure 7) and therefore does not appear to be affected by early burial diagenesis. During deeper burial, the δ 18 O signature might be modified by further cementation and recrystallization potentially overprinting the primary isotopic differences (Reuning, Reijmer, & Betzler, 2002), but the δ 13 C signature is expected to be less prone to alteration (Reuning, Reijmer, & Mattioli, 2006;Reuning et al., 2002). Besides the planktic component, the isotopic signature therefore seems to be a criterion that can be used to differentiate shallow water tropical pellets from those deposited in deeper, and therefore cooler, waters.
early cementation of pellets in this area. A fossil example for the importance of very early cementation in the consolidation of faecal pellets is given by Friis (1995). He demonstrated that faecal pellets were only preserved in Tertiary siliciclastic sediments if they were stabilized by very early pyrite and calcite cements. Similarly, it is suggested here that very early cementation by calcite, dolomite and pyrite related to aragonite dissolution and BSR were essential for the preservation of the faecal pellets after burial.

| CONCLUSIONS
Faecal pellets can be as abundant in deeper water, subtropical environments as in shallow, tropical settings. On the Carnarvon ramp, faecal pellets are formed and preserved at the transition between the outer ramp/shelf and slope at a depth >200-300 m. They occur more abundantly during interglacials, when IODP site U1460 was situated in deeper waters below the swell wave base. Faecal pellets from deeper, cool-water carbonate settings are characterized by lower δ 13 C values and higher δ 18 O, and by higher planktic components compared to shallow water tropical peloids. Furthermore, the fact that they are composed of planktic and benthic components differentiates them from pellets produced by pelagic organisms. The near sea floor cementation by calcite, dolomite and pyrite related to BSR-driven aragonite dissolution is essential for the hardening of faecal pellets and their preservation in the fossil record.