A new macrofaunal limit in the deep biosphere revealed by extreme burrow depths in ancient sediments

Macrofauna is known to inhabit the top few 10s cm of marine sediments, with rare burrows up to two metres below the seabed. Here, we provide evidence from deep-water Permian strata for a previously unrecognised habitat up to at least 8 metres below the sediment-water interface. Infaunal organisms exploited networks of forcibly injected sand below the seabed, forming living traces and reworking sediment. This is the first record that shows sediment injections are responsible for hosting macrofaunal life metres below the contemporaneous seabed. In addition, given the widespread occurrence of thick sandy successions that accumulate in deep-water settings, macrofauna living in the deep biosphere are likely much more prevalent than considered previously. These findings should influence future sampling strategies to better constrain the depth range of infaunal animals living in modern deep-sea sands. One Sentence Summary: The living depth of infaunal macrofauna is shown to reach at least 8 metres in new habitats associated with sand injections.


Geological setting
The Permian Ecca Group, a succession of siliciclastic material, was deposited within the Tanqua and Laingsburg depocentres of the SW Karoo Basin [31]. The Laingsburg depocentre (Fig. 1) comprises a shallowing upward stratigraphic succession. This study focuses on outcrops of the Fort Brown Formation; a 400 m thick submarine slope succession [31,32] (Fig. S1). The Fort Brown Formation comprises sand-prone lithostratigraphic units C to G, which are further divided into subunits by laterally extensively thin (<2 m thick) fine siltstones consistent with palaeogeographic changes across a unit [12,33]. Each unit is separated by a thick regionally extensive mudstone (siltstone and claystone) (>10 m thick) interpreted to represent a basin-wide shut down in sand supply [31,32].
Small burrows occur within mudstone beds that separate sand beds [11], in thin-bedded siltstones [13] and on the base of structureless and rippled sandstones [11]. Ichnofacies assemblages are interpreted to be primarily Chondrites and Planolites [11].
Individual outcrop sites are indicated by place names and stratal units (Figs. 1 and S1).
At each site, the units are submarine lobe deposits [12,13] and comprise thin, very fine sandstones capped by mudstones ( Fig. 3A and 3C). Clastic injectites are recognised through dykes cross-cutting stratigraphy, sills stepping up and down stratigraphy, and their sharp sided nature on top and base margins [28,29]. In the Karoo Basin, injectites are sourced from the very fine sandstone units, they are up to 0.5 m thick, sharp sided, and usually subvertical. At sites 1, 2 and 3 the injectites are 8 m, 1.5 m and 3 m compacted depths respectively, below the capping sand. The same trace fossils present on the base of units (Fig. 2B) are also observed on the margins of clastic injectites down to their lowermost occurrence (Fig. 3).

Outcrop observations
Site 1: Unit E, Geelbeck At Geelbeck, Subunit E1 is absent, Subunit E2 is an intraslope lobe complex comprising three stacked lobe deposits, the lowermost of which was deposited in a highly-confined environment [13]. The base of Subunit E2 is very well exposed, and therefore so are examples of Planolites (Fig. 2); individual borrows are up to 10 cm long and <1 cm in width. The basal beds of Subunit E2 onlap an erosional surface and mainly comprise structureless sandstones, with ripples present towards the bed tops.
The dykes directly in contact with the base of Subunit E2 form an abruptly downward tapering or upward opening cone ~1 m in width and 1.5 m in depth (Fig. 2C). Each conelike feature comprises multiple dykes passing into or from a single dyke 5-20 cm wide that are connected to multiple sills and dykes below. Sills do not step through stratigraphy and individually extend up to 20 m. On the margins of the clastic injectites, the same trace fossil, Planolites, is present (Figs. 3B-3F) that occurs on the base of Subunit E2 (Fig. 2B). Additionally, on a subvertical dyke ~2 m below the base of Subunit E2 are dewatering structures (Aristophycus) (Fig. 3A), that are overprinted by Planolites.
The bioturbation is present on tops and bases of sills and on the margins of dykes and is observed at least 8 m (compacted thickness) stratigraphically below Subunit E2, which is the extent of the injectites.
Site 2: Unit D, Slagtersfontein West Unit D at Slagtersfontein West is represented by the lowermost sequence, Subunit D1.
Subunit D1 is interpreted to be lowstand lobe deposits [12] similar in character to those described by [34]. Where observed, Subunit D1 has a sharp, sometimes erosive, laterally extensive basal contact with the underlying mudstone unit and comprises structureless sand. The subunit is locally deeply incised by a younger submarine channel surface [35].
Injectites are observed directly below the base of Subunit D1 as <40 cm thick dykes, which penetrate at least 2 m of the underlying mudstone (Fig. S2B). The margins of these dykes exhibit randomly orientated, and up to 20 cm long, Planolites (Fig. S2B).

Site 3: Unit C, Slagtersfontein East
Here, C2 is the only subunit of Unit C present and is interpreted to be the proximal edge of a lobe complex [12]. Locally, the basal surface of Subunit C2 cuts into the underlying B/C mudstone with the lower beds consisting of structureless and amalgamated sandstones. Planolites is present along the base of Subunit C2. Injectites exposed in the Slagtersfontein area are primarily hosted within the regional mudstone separating Units B and C (Fig. S2A). The majority of injectites at the Slagtersfontein outcrop are 0.1-0.6 m thick sills that extend laterally for up to 500 m. These connect to subvertical dykes directly in contact with the base of Subunit C2. Planolites is present on the top and base of sills and dykes up to 2 m stratigraphically below C2. The burrows are randomly orientated and common across the injectites.

Outcrop summary
In each of the cases presented here, clastic injectites are hosted within mudstones, and there is a direct connection between the injectites and the overlying capping sandstones. The grain-size is too narrow and provenance too consistent in the sandstone of the Fort Brown Formation to be used to discriminate the source of the injectites. Furthermore, two-dimensional outcrops make it difficult to determine if the sands were injected from above or below since a deeper connection may be out of the plane of the section. In all cases, there is evidence for erosion on the surface between the capping sands and the underlying mudrock that hosts the injectites. Given this, and the thin and silty nature of the overlying fine-grained units that suggest that the lobe sands were not rapidly buried by mud-rich flows, the scenario where the injectite networks existed prior to exploitation by macrofauna appears more likely. However, we cannot discriminate whether the injectites were sourced from below and connected to the seabed to form extrudites, or were exhumed through submarine erosion with an underlying or overlying source, although an underlying parent sand is considered more likely given pressure gradients and depth of erosion required.

Model of oxygen consumption in injected sediments
The first scenario presented assumes no replenishment of oxygen and POM (particulate organic matter), therefore it is necessary to calculate the amount of time macrofauna Pacific [37]. For the initial oxygen concentration, we used a value of 8 ml / L reflecting relatively well-oxygenated water overlying seafloor sediments. We assume the site is not in an OMZ. To correct for inclusion of sandy sediment in the volumetric space, we assume a porosity (Φ) of 46% such that: This is typical of the sediment injected, fine to very fine sands.
All the biologic activity within the sediment is accounted for by the sediment community oxygen consumption (SCOC). This includes macrofauna, however, they typically only account for a small portion of the SCOC, whereas in the model presented, we add an additional polychaete population. [38] measured values of total oxygen uptake in sediments ranging from 0.0403 -0.347 mlO2L -1 day -1 in the south east Atlantic. They suggest these are higher than typical values elsewhere due to high surface productivity.
[39] measured median value in sediments east of Svalbard, of 0.0618 mlO2L -1 day -1 . An average presented by [40] for the north Atlantic and Pacific from depths of between 1 and 2 km is 0.0508 mlO2L -1 day -1 . Here, we used the median value of 0.0618 mlO2L -1 day -1 , since this included sampling from sandy sediments, at high latitude, more analogous to the deposits in the Karoo Basin.
Samples from the Rockall Trough (West of Scotland) found an average of 0.39 -1.724L -1 depending on the mesh size used to sort sediments [40]. Whereas [41] found macrofauna abundances of around 1.8 L -1 . We used a value of 2 L -1 .
Metabolism of deep-sea organisms scales in a similar way to shallow water species, where size and temperature account for most of the variability [21]. [20] empirically showed that the respiration rates of deep sea organisms (taken from areas of 2-4°C) scales with their weight such that: where R is the respiration rate (per day) and W is the weight (in mgC) of the organism.
We modified this equation to give the respiration rate in units of ml02day -1 such that: where 1/0.44 is the mobilisation of oxygen (in ml) per mg of carbon (taken from [20].
The weight used was 0.428 mgC, the average size for the deep sea macrofauna used by [20] in their study of nematodes, copepods and polychaetes.
As data were unavailable, a value of 0.5 was used, i.e. only half the population of polychaetes survive the injection (or actively colonize the injectite) to exploit the injectite network post-emplacement.

Sb
As data were unavailable, a value of 0.5 was used. Mechanical shaking of the sediments causes microfauna to be lost as they typically have lower densities than sediments.
Further, some proportion of reduced chemicals in the sediment will be oxidised as mixing with overlying waters occurs, reducing oxygen uptake by chemical means postinjection.

Unit Conversions and results:
Units reported in the literature needed to be converted in many cases prior to being input into the model. In the literature, SCOC and abundances are typically reported as per unit area of sediment surface, therefore these have been converted to volume to provide a depth aspect on oxygen consumption within a community by assuming a depth of 1 m. Molar oxygen concentrations were converted to ml / L using the ratio of 1 mlO2 / L seawater = 44.661μmolO2 / L (from ICES oceanography).
Below an oxygen concentration of 0.45 ml / L, the community structure of deep-sea macrofauna becomes adversely affected, however it appears that polychaetes are the most tolerant of macrofaunal taxa [36]. We therefore took the threshold of polychaetes to be 0.2 mlO2 / L. A length of time can therefore be estimated, before oxygen becomes too low in the sediment. As parameterised above, this occurs at 269.8 days, which provides time for burial and sealing by mud-rich turbidity currents and injection to occur (Scenario 1), or, for macrofauna to exploit injectites and travel down to their lower most occurrence (Scenario 2 and 3) (Fig. S3).
An analytical sensitivity analysis of the model is shown in Fig. S4, and demonstrates that the model is relatively insensitive to the precise parameters used, and consequently there would have been sufficient time for organisms to have produced the observed burrows.