Meiofaunal traces in cored Pleistocene to Holocene silt-clay glaciogenic rhythmites consist of four types: clay-rich fill (type A) representing feeding structures, and a silt-rich fill with a clay-rich lining (type B) representing dwelling structures. A third burrow type (type C) is filled with framboidal pyrite suggestive of microbially mediated early diagenesis under anoxic and circumneutral pH conditions. A variant of type C is filled with framboidal pyrite overgrown by poorly crystalline pyrite indicating at least two stages of iron sulphide growth within earlier formed burrows. Higher in the cored succession a bed thickness increase to about ∼ 14 cm is marked by an absence of meiofaunal traces. Since each bed represents deposition during a single year, it suggests that sustained sedimentation rates of ∼ 14 cm yr−1 represents a threshold for endobenthic activity. A fourth burrow type (type D) is filled mostly with clay and a lining with a distinctive parallel-aligned fabric of clay and dispersed silt grains, which occurs in slump blocks resedimented downslope from a shallower water setting. The absence of type A, B, or C burrows in these silt-rich slump blocks suggests that prior to slope failure, the silt-rich sediment substrate, and likely brackish pore fluid conditions, were not conducive to these tracemakers and/or trace-making behaviors. These differences illustrate the first-order control of physico-chemical factors, specifically, sedimentation rate, salinity, and substrate composition on the behavior of burrowing meiofauna in a glaciomarine basin plus the significance of meiobenthic activity on controlling early, shallow burial diagenesis.

Ichnological studies typically focus on the burrowing and feeding traces of macrofauna (i.e., > 1 mm in diameter; Andersen and Pejrup 2011), which in part can be attributed to the relative ease of recognizing and describing visibly conspicuous macrobenthic trace fossils in core and outcrop (e.g., Bromley 1996; MacEachern et al. 2005; Buatois and Mángano 2011). In contrast, meiofauna, which include small invertebrate animals like nematodes, foraminifera, and copepods (Baguley et al. 2019), and their associated traces (< 1 to ∼ 0.045 mm in diameter; Giere 2009), although long recognized by marine biologists as a ubiquitous component in modern marine ecosystems (e.g., Mare 1942; Gerlach 1971; Warwick 1989), have been much less documented in the sedimentary record. This may be partly due to their small size but potentially also the negligible compositional and/or textural contrast, and therefore optical contrast, between the burrow fill and the host (surrounding) sediment (Wilson et al. 2021). However, meiofauna occupy a diverse range of sedimentary environments, from lakes to the deep ocean (Giere 2009), and considering their much greater numerical abundance in comparison to macrobenthos (e.g., Wigley and McIntyre 1964), are significant contributors to sediment bioturbation (e.g., Cullen 1973; Brodie and Kemp 1995; Pike et al. 2001; Pemberton et al. 2008; Lӧhr and Kennedy 2015; Paz et al. 2023). Despite their small size, meiobenthic burrowers commonly line their burrows with mucous (organic matter) that later becomes susceptible to microbial degradation and the formation of early diagenetic minerals that help enhance their visibility, but more importantly, provide insight into paleo-redox conditions, element cycling, and mineral-forming processes through geologic time (e.g., Virtasalo et al. 2010; Gingras et al. 2014; Harazim et al. 2020).

Recent advances in imaging techniques, like high-resolution optical and electron microscopy and X-ray imaging, have contributed to our understanding of the occurrence and distribution of meiofaunal traces in the sedimentary record (Parry et al. 2017), and the effect that meiofauna have on primary sedimentary fabric (e.g., Biddle et al. 2021; Schieber and Wilson 2021). Most of these studies have focused on modern and ancient open marine environments, whereas little attention has focused on the deposits in glaciolacustrine and glaciomarine settings. An understanding of the stratigraphic record in glaciogenic settings is significant as they contain the sedimentary archive that record the evolution of glacial conditions, specifically, physical, chemical, and biological processes that operate within them (Domack and Powell 2018; Fitzsimons and Howarth 2018). The ichnological suites of glaciogenic settings, like open marine settings, are temporally and spatially influenced by variable physico-chemical stresses (Buatois et al. 2006; MacEachern et al. 2007; Buatois et al. 2010; Gingras et al. 2011; Alonso-Muruaga et al. 2013; Schatz et al. 2013), and in the case of the glaciomarine realm, these prevalent stresses are often sedimentation rate and fluctuations in salinity (e.g., Cowan et al. 1999; Eyles and Eyles 2010). Typically, the influence of these physico-chemical controls on an ichnofabric are challenging to determine, especially when the absolute duration of sedimentation is unknown. However, in a recent sedimentological investigation of Late Pleistocene to Holocene glaciolacustrine to glaciomarine deposits, Al-Mufti et al. (2022) attributed all beds to reflect the annual glaciogenic waxing followed by waning of glacial meltwater discharge, thereby providing a temporal context to examine the nature of bioturbation and its relationship with a number of paleoenvironmental controls, like sedimentation rate and salinity, in a glaciogenic setting.

In this study, glaciolacustrine to glaciomarine strata of the Late Pleistocene to Holocene Champlain Sea basin provide a nearly modern analogue to examine the occurrence and distribution of meiofaunal traces and associated diagenetic minerals in subaqueous glaciogenic mud that has experienced shallow burial (< 66 m). More specifically, the objectives of this study are to (1) determine the detrital grain mineralogy and describe the composition, texture, and fabric of biogenic structures and diagenetic minerals in glaciogenic mud; (2) interpret the origin of the different types of biogenic structures and assess organism-sediment relationships; (3) determine the origin of diagenetic minerals, their spatial and temporal association with biologic activity, and infer post-depositional redox conditions; and (4) discuss paleoenvironmental controls on meiobenthos in the Champlain Sea basin.

Champlain Sea Basin

The Laurentide Ice Sheet (LIS) was the largest Pleistocene ephemeral ice sheet in the Northern Hemisphere, and at its maximum extent covered much of the surface area of present-day Canada and the northeastern United States (Dyke et al. 2003; Margold et al. 2015). Northward retreat of the ice front from its maximum extent during the late Wisconsinan glaciation (18 ka BP; Dyke et al. 2003) resulted in inundation of the isostatically depressed Lake Champlain, Ottawa River, and St. Lawrence River valleys (Cronin 1977; Fulton et al. 1987) by freshwater glacial Lake Candona (Parent and Occhietti 1988). Continued retreat of the ice front northward of Québec City resulted in the rapid incursion of Atlantic Ocean seawater into the glacio-isostatically depressed basin (Cronin 1977) and converted Lake Candona in as little as several years (Al-Mufti et al. 2022) into the Champlain Sea (Fig. 1A). The Ottawa area of the Champlain Sea basin served as a major depocenter for the accumulation of an up to ∼ 100 m-thick succession of glaciolacustrine (Lake Candona) overlain by glaciomarine (Champlain Sea) mud (Gadd 1987; Medioli et al. 2012); although, most of the stratigraphic column corresponds to the glaciomarine phase of sedimentation (Al-Mufti et al. 2022). Collectively and historically this mud is commonly referred to as ‘Leda clay’ or ‘Champlain Sea mud/sediment’ and has been the focus of numerous studies for more than a century (e.g., Johnston 1917; Elson 1988 and references therein; Cummings et al. 2011; Al-Mufti et al. 2022).

In a recent study of sediment cores collected from the Champlain Sea basin, Al-Mufti et al. (2022) used X-ray computed tomography (CT) to identify five lithologically distinct, mud-dominated stratal units and their macroscopic (> ∼ 0.5 mm-scale) characteristics (Fig. 2). From base to top, these units are laminated mud rhythmites (Unit 1), bioturbated mud (Unit 2), banded mud (Unit 3), well-stratified mud interstratified with diffusely stratified or structureless mud (Unit 4a), and deformed mud (Unit 4b). By analyzing quantitative CT-scan data and images, Al-Mufti et al. (2022) showed that all beds within these units are characterized by a progressive upward increase followed by decrease in radiodensity, reflecting rhythmic changes in bulk density caused by differences in silt abundance, and therein recording the waxing followed by waning of glaciogenic meltwater discharge. Moreover, the consistent rhythmicity of bed-scale variations in silt/clay content was taken to reflect seasonal variations in glacial meltwater discharge, with each bed attributed to deposition over a single year of sediment accumulation, and accordingly, considered to be a glacial rhythmite or ‘varve’ (e.g., Ashley 1975; Ridge et al. 2012).

Study Area

The study area is in a suburban area in Ottawa, Ontario, Canada (Fig. 1B). In autumn 2019, two boreholes, separated by ∼ 3 km, were drilled along Voyageur (VC2) and Bilberry (BC16) creeks, which are tributaries that flow northward into the Ottawa River. Borehole VC2 was drilled to a depth of 66 m and BC16 to 49 m (Fig. 3). Nearly continuous sediment cores were collected using 76.2 cm-long Shelby sampling tubes with a nominal diameter of 7.3 cm. Drilling at VC2 was terminated in Unit 2 before intersecting Unit 1 and the underlying glacial till, whereas BC16 was cored to the contact with the till (Hinton and Alpay 2022). Fine-grained (clay- to silt-size) deposits sampled from these boreholes provided the opportunity to document the microtextural characteristics of the sediment.

Sediment cores collected from the boreholes were sealed on-site and preserved upright in a 3–5°C cold storage at the Geological Survey of Canada (Ottawa, ON) (Alpay et al. 2020). Although the cores consist of unlithified sediment buried up to only a few tens of meters (< 66 m), the water-saturated, fine-grained nature of the sediment rendered them relatively resistant to breakage during extrusion and subsampling. Cores were extruded from their steel tube casings and ∼ 5 cm-long sections were cut carefully from the cores and then placed in a metal container thinly lined with non-combustible insulation. A lamp with a 60-watt incandescent light bulb was placed on its side and positioned in the metal container wrapped with insulation with a small hole at the opposite end of the container to allow any water vapor to escape. This ‘oven’ resulted in an internal temperature of about 45°C, and samples were then left to dry for one week. The completely desiccated samples were then impregnated with an epoxy resin (EpoFix®, Electron Microscopy Sciences, PA, USA) heated to a temperature of 50°C to lower viscosity and improve penetration. Next, resin was poured slowly over the samples and left to cure for four days, followed by a second coat of resin to an adjacent perpendicular face. The samples were then cut into rectangular slabs using a diamond saw with no water lubrication to minimize clay hydration and sample distortion. The cut slabs were made into regular thin section slides and polished down to ∼ 25 μm using a 0.25 μm diamond grit plate at the University of Western Ontario (London, ON). In total, 35 polished thin sections were produced from the boreholes (29 thin sections from VC2 and six thin sections from BC16), and each thin section was scanned at 2400 dpi using a high-resolution flatbed scanner. Thin sections were then examined petrographically using an Olympus BX-41 petrographic microscope and later coated with gold-palladium and analyzed using a JEOL-6610LV scanning electron microscope at 20 kV under high vacuum and at a working distance of 10 mm. The chemical composition of unknown minerals was analyzed using energy dispersive X-ray spectroscopy. Ichnologic analyses were conducted on all thin sections, including description of ichnotaxobases, burrow size, bioturbation index, and interpretation of ethology and potential tracemakers. Characterization of ichnotaxobases included the general shape and orientation of burrows, details of the burrow boundary (wall or lining), whether burrows are branched or unbranched, and the composition of the fill. All thin sections were cut perpendicular to bedding allowing burrow diameter to be measured directly whereas morphological and dimensional data measured in the plane-of-bedding were obtained from transverse (bedding-parallel) X-ray computed tomography images of the sediment cores. The approximate location of all captured optical and scanning electron micrographs in the thin section scans is provided in the Online Supplemental File 1.

Powder X-ray diffraction (pXRD) analysis was completed on seven sediment samples to determine the semi-quantitative mineralogy (wt. %). For five samples, pXRD was completed on the mixed silt- and clay-size fraction, and for two samples, pXRD was completed on the < 2 μm and > 2 μm size fractions. Grain size separation was completed at the Geological Survey of Canada (Ottawa, ON). X-ray patterns of the powder samples were recorded at the University of Ottawa on a Rigaku Ultima IV X-ray diffractometer equipped with a scintillation counter, and a sealed-tube Cu Kα source set at 40 kV and 44 mA. Initial identification of minerals was made using the HighScore Plus (Malvern Panalytical) software with comparison to reference mineral patterns using the PDF-4+ database from the International Centre for Diffraction Data (ICDD). Quantitative analysis by Rietveld refinement was also carried out using HighScore Plus. Weight percentage data is provided in Online Supplemental File 2.

Lithology, Detrital Assemblage, and Bed Thickness

Microscopic examination (< 0.1 mm-scale) of lithostratigraphic units 1 through 4b reveals that sediment is composed mostly of clay- to coarse silt-size grains of quartz, plagioclase feldspar, potassium feldspar, biotite, and hornblende. Coarse silt to fine sand-size foraminifera are dispersed in bioturbated mud of Unit 2. Bulk pXRD analysis (semi-quantitative; Table 1) of the silt- and clay-size fraction of sediment from units 1 through 4b indicates that the mineral assemblage is composed of quartz (∼ 13–20%), albite (∼ 17–35%), microcline (∼ 9–33%), and biotite (∼ 9–22%), with lesser amounts of hornblende (∼ 4–12%), chlorite (∼ 3–27%), and trace amounts of calcite. Notably, and irrespective of stratigraphic position, this mineralogy is consistent throughout the succession. The pXRD analysis of the clay-size fraction also reveals a nearly identical mineralogy to the bulk silt-clay-size fraction (Table 1). Bed thickness, although variable, shows a general upward thickening and increases abruptly across unit boundaries (Fig. 4). More specifically, bed thickness averages: Unit 1: ∼ 2 cm, Unit 2: ∼ 4 to 6 cm, Unit 3: ∼ 10 cm (lower part) to ∼ 14 cm (upper part), and Unit 4a (∼ 20 cm).

Biogenic Structures

Four distinct burrow morphologies are recognized in Champlain Sea mud: burrows composed of a clay-rich fill (Type A), burrows composed of a silt-rich fill with a clay-rich lining (Type B), burrows filled with iron sulphide (Type C), and burrows composed of a clay-rich fill with a clay-rich lining (Type D). Type A, B, and C burrows occur in beds of bioturbated mud (Unit 2) and the lower part of banded mud (Unit 3). Type D burrows occur exclusively in deformed mud (Unit 4b). The degree of bioturbation is described using the bioturbation index of Taylor and Goldring (1993), where a value of 0 indicates no bioturbation (Unit 1), 4 to 5—high to intense bioturbation (Unit 2), 1 to 2—sparse to low bioturbation (lower part of Unit 3), and 0 for the upper part of Unit 3 and all of Unit 4a.

Type A burrows are unbranched and unlined horizontal structures typically filled with clay-size sediment and readily differentiated from the surrounding mud matrix (Fig. 5A, 5B). The outer margin of the burrow fill is smooth and the transition from the burrow fill to the burrow margin is sharp (Fig. 5C). Type A burrows are ∼ 500 to 870 μm in diameter and in cross-section exhibit either an elliptical (Fig. 5B, 5D, 5E), sub-circular (Fig. 5C), or irregular (Fig. 5F) shape. Although type A burrows lack a distinct lining, some burrow margins are marked by an up to 100 μm-thick layer that is slightly more enriched in silt relative to both the burrow fill and the surrounding mud matrix (Fig. 5E, 5F).

Type B burrows are unbranched horizontal structures consisting of a silt-rich fill lined with clay-size sediment (Fig. 6A). The silt-rich fill ranges from ∼ 130 to 240 μm in diameter and consists mostly of very fine to medium silt with uncommon coarse silt-size grains (Fig. 6B). Clay-rich linings range from ∼ 20 to 70 μm thick and consist mostly of clay-size sediment with dispersed very fine to fine silt (Fig. 6C). Linings of individual burrows are either nearly isopachous or vary irregularly around the burrow by up to ∼ 50 μm (Fig. 6D). Moreover, burrows can be easily differentiated from the mud matrix by the relative enrichment of silt grains in the fill and the optically distinct birefringence of the clay-rich lining (Fig. 6E). In cross-section, most type B burrows are elliptical to sub-circular in morphology, and in transverse views, exhibit a sinuous to elongated planform (Fig. 6F).

Type C burrows are ∼ 70 to 190 μm wide and up to 3000 μm long. Burrows are horizontal and filled completely with various forms of iron sulphide, namely, clusters of pyrite (FeS2) framboids, framboidal pyrite overgrown by poorly crystalline pyrite, or semi-elliptical concretionary masses of iron sulphide. Type C burrows do not appear to exhibit any preferential grain size segregation. See ‘Diagenetic Mineralogy’ (next) for a detailed description of these cement-filled burrows.

Type D burrows are horizontal structures filled with structureless clay and lined with clay to fine silt (Fig. 7A, 7B). Burrow fills range from ∼ 40 to 80 μm in diameter and their linings from ∼ 20 to 50 μm in diameter. They occur as a closely packed assemblage of similarly shaped and sized burrows (Fig. 7A, 7B). In comparison to their lining and the adjacent matrix, most burrows show a greater abundance of clay-size particles in the center of their fill (Fig. 7C7E). Distinctively, type D burrow linings show a well-defined fabric consisting of non-grain-size-segregated clay- to fine silt-size particles that are aligned parallel to the burrow boundary (Fig. 7F).

Diagenetic Mineralogy

Microscopic examination shows iron sulphide to occur as: (1) elongated clusters of pyrite framboids; (2) semi-elliptical concretionary masses of indiscrete crystals; (3) sub-micron-size crystals between the cleavage sheets of mica grains; and (4) pyrite framboids overgrown by poorly crystalline pyrite. Pyrite framboids form clusters with mean diameters ranging from ∼ 5 to 15 μm and composed of sub-micrometer-scale cubic microcrystals and intercrystalline porosity (Fig. 8A, 8B). Individual framboids within clusters are surrounded by innumerable, loosely packed cubic microcrystals (Fig. 8C). Clustered framboidal pyrite fills elongated cavities that are ∼ 150 to 470 μm long (Fig. 8A, 8D) in bedding-perpendicular sections and up to 3000 μm long in transverse sections (Fig. 8E). Semi-elliptical concretionary masses of iron sulphide are composed of sub-micrometer-scale crystals with a poorly developed crystalline habit (Fig. 9A, 9B). Moreover, these concretionary masses are ∼ 60 to 100 μm in diameter with high intercrystalline porosity. Dispersed clay- to fine silt-size detrital quartz, feldspar, and micas are contained within these concretionary masses (Fig. 9B) and EDS analysis of the iron sulphide reveals traces of Si, Al, Na, and Ca. Additionally, EDS analysis shows the intensity of the sulphur peak is about half that of pyrite. Subhedral to euhedral iron sulphide crystals ≤ 1 µm in diameter occur in the space between the expanded platelets of some mica grains. The extent of this style of iron sulphide cementation ranges from locally developed, where crystals occur either discretely or in clusters (Fig. 9C), to extensive, where the interstitial pore space between cleavage sheets is filled completely with crystalline iron sulphide (Fig. 9D).

Some pyrite framboids are overgrown with poorly crystalline pyrite (Fig. 10A). This form of pyrite is characterized by crystals with a tabular-bladed morphology (Fig. 10B), and like the clustered occurrences of framboidal pyrite (Fig. 8), typically fill elongated and meandering cavities (Fig. 10C10E). Dispersed detrital clay- to silt-size minerals occur in pyrite overgrowths (Fig. 10D10F). Clusters of framboidal pyrite and forms of pyrite framboids overgrown by poorly crystalline pyrite occur in approximately sub-equal amounts.

Pyrite is absent in laminated mud rhythmites in the lowermost part of the succession (Unit 1) and first appears in a thin (∼ 25 cm) stratigraphic interval that marks the upward transition to bioturbated mud (Unit 2). All texturally distinct forms of iron sulphide are observed in bioturbated mud of Unit 2 and the lower part of the distinctively diffusely banded mud of Unit 3 but are absent in the upper part of Unit 3 and overlying well-stratified (Unit 4a) or deformed mud (Unit 4b) in the upper part of the succession.

Origin of Biogenic Structures

Four burrow types (A–D) were identified based on their morphological and textural characteristics, including their overall morphology, orientation, and composition, texture, and fabric of their fill and lining (or lack thereof). The small size of the burrows (several tens to hundreds of (µm) and their sinuous to elongated planform (e.g., Figs. 6F, 8E) resemble structures formed by modern marine meiofauna (e.g., Jensen 1996; Giere 2009). Notably, a lack of centimeter-scale burrows and shells indicates a conspicuous scarcity to absence of burrowing macrofauna.

Type A burrows consist of horizontal unbranched burrows with a clay-rich fill and are interpreted to represent deposit-feeding structures. These unlined traces are characterized by burrow fills that are readily differentiated from the surrounding substrate and resemble Planolites-like traces (e.g., Pemberton and Frey 1982). Moreover, these burrows are similar to ‘Planolites type A’ burrows described by Egenhoff and Fishman (2013); however, unlike the lenticular forms (0.3 to 1 mm across and < 0.1 mm in height) in the Upper Devonian–Lower Mississippian mudstones described by Egenhoff and Fishman (2013), type A burrows reported here occur in Quaternary mud that has undergone negligible compaction (Al-Mufti et al. 2022) and therefore are circular to slightly elliptical in cross-section. Planolites is interpreted to represent the backfilling with sediment of ephemeral (transient) burrows constructed by mobile deposit-feeding, infaunal organisms (Pemberton and Frey 1982). The absence of a well-developed lining is typical of mobile organisms that do not construct dwellings, but rather exhibit grain-selective feeding behaviors (Gingras et al. 2011). Moreover, the presence of silt grains along the fill-margin boundary (Fig. 5E, 5F) might reflect grain-selective feeding or are relict silt grains that became concentrated along the margins as the organism tunneled through the sediment (Biddle et al. 2021). Planolites filled with sediment finer than the surrounding substrate (e.g., Frey and Bromley 1985), like type A burrows in this study, are much less common than those filled with sediment texturally coarser and more resistant than the substrate matrix (e.g., Pemberton and Frey 1982; Dafoe et al. 2010; Zonneveld et al. 2010). The sub-millimeter-size of the Planolites-like burrows necessitates burrowing by meiobenthic organisms, which might include worm-like organisms (e.g., Harazim et al. 2020) or carnivorous and scavenging polychaetes (e.g., Naldrett 1990; Gingras et al. 2011).

Given their distinct and well-developed clay-rich linings, type B burrows are interpreted to be infaunal dwelling structures. Organisms that produced these burrows may have actively lined their tunnels with a layer of clay to stabilize the substrate and provide mechanical strength to the burrow wall (Gingras et al. 2011). Additionally, the burrow lining may also serve to maintain an open conduit to facilitate irrigation, and/or if lined with mucous, provide lubrication for ease of movement of infaunal tunnellers (Gingras et al. 2011). The enrichment of silt in type B burrow fills compared to the surrounding mud matrix is interpreted to represent the passive filling of the open burrow, and therein resembles forms of the ichnogenus Palaeophycus (e.g., Pemberton and Frey 1982). Moreover, the abundance of silt-size grains in the fill compared to the surrounding mud matrix, suggests that sediment was most probably sourced from silt-rich deposits (e.g., laminae), indicating finer sediment texture prior to homogenization by infauna. The passive filling of type B burrows contrasts with the active filling origin of type A burrows by the related ichnogenus Planolites (Pemberton and Frey 1982) and indicates differences in meiobenthic tracemakers and/or trace-making behaviors. Possible meiobenthic tracemakers that may have produced these infaunal dwelling structures include mud dwelling nematodes (Nehring et al. 1990; Giere 2009) and predaceous polychaetes (Naldrett 1990).

See ‘Origin of Iron Sulphide’ (below) for an interpretation of type C (cement-filled) burrows.

Type D burrows consist of a central clay-rich core surrounded by a lining of clay- to fine silt-size particles. The well-defined fabric of the burrow lining is characterized by clay- to fine silt-size particles aligned parallel to the tunnel margin. Accordingly, the fabric of type D burrow linings is interpreted to be the result of circular body contractions by a meiofaunal organism that compacted undigested sediment against the burrow boundary (Zorn et al. 2010). More specifically, sediment compaction against its burrow margins may have been achieved during animal burrowing by peristaltic contractions, as in some annelids, or by the extension-retraction of an introvert (e.g., loriciferans, priapulids, sipunculids; Powilleit et al. 1994; Giere 2009; Parry et al. 2017).

Origin of Iron Sulphide

Clusters of framboidal pyrite fill elongated structures that morphologically resemble burrows formed by burrowing meiofaunal animals (e.g., Virtasalo et al. 2010; Parry et al. 2017). The open fabric of the pyrite framboids within the type C burrows indicates that they formed prior to compaction, most probably in voids that were not passively filled as in type B burrows, or grain-size-segregated as in type A and D burrows. Moreover, framboidal pyrite is considered to have precipitated during early diagenesis when porewater sulphide from organic matter decomposition reacted with dissolved ferrous iron to produce a series of metastable iron monosulphides, like mackinawite (Berner 1984), that then transformed to pyrite through further reaction with sulphur (Raiswell 1982). In bioturbated sediment, burrows often contain organic matter in the form of ‘slime’, ‘mucus’, or the more general term ‘extracellular polymeric substances’ (EPS; Flemming 2011), which meiofauna often use to line their burrows to facilitate movement (lubrication) through the sediment (Bromley 1996; Gingras et al. 2011). This kind of organic matter is easily metabolized and was most probably the principal bioavailable reductant that was oxidized during microbially mediated sulphate reduction (Poulton et al. 1998). Additional sources of organic matter may have included fecal material of infaunal burrowers or decomposition of the tracemaker itself in the burrow (Wilson and Brett 2013). The preferential occurrence of pyrite framboids within the burrows and their general absence in the surrounding sediment necessitates the requisite role that reactive organic matter (EPS) served as a bioavailable reductant for microbial sulphate reduction (Berner 1985; Berg et al. 2020).

In comparison to the clustered forms of framboidal pyrite, semi-elliptical concretionary iron sulphide is interpreted to have precipitated in sediment less enriched in organic matter (e.g., Virtasalo et al. 2010, 2013). In these sediments the lower abundance of reactive and easily metabolized organic matter resulted in semi-elliptical, polycrystalline masses of iron sulphide cement. EDS analysis of these crystals shows the intensity of the sulphur peak to be about half that of pyrite, suggesting that these crystals may be an intermediate precursor to pyrite, like mackinawite or greigite (e.g., Schieber 2011a, fig. 1). Traces of Si, Al, Na, and Ca, in addition to clay-size sediment surrounding the crystals (e.g., Fig. 9B), indicates that this form of iron sulphide admixed with the surrounding detrital matrix (clay) as it precipitated.

Many of the clustered pyrite framboids are overgrown by poorly crystalline laths of pyrite, indicating that it postdates framboidal pyrite formation. The abrupt change from framboidal pyrite to poorly crystalline pyrite reflects an abrupt reduction in the rate of precipitation under low hydrogen sulphide concentration (Virtasalo et al. 2013). Moreover, the change in crystalline form may be due to slower microbial metabolism of the remnants of less labile organic matter (EPS) along burrow walls. Although the abrupt change in pyrite type suggests a change in precipitation rate, the widespread extent of these pore-filling crystals indicates that, at least locally, sulphate-reducing bacteria continued to create chemical conditions conducive to pyrite formation under reducing conditions with a source of dissolved ferrous iron (Fe2+) and H2S. As poorly crystalline pyrite grew, the pre-formed pyrite framboids formed a nucleation surface for continued iron disulphide growth until the burrow filled completely with pore-filling pyrite.

Timing of Cementation

Each bed in the sedimentary succession has previously been interpreted to represent deposition during an annual cycle of glaciogenic meltwater discharge, and therefore are collectively considered to be glacial rhythmites, or ‘varves’ (Al-Mufti et al. 2022). Burrows containing framboidal pyrite (type C), in addition to type A and B burrows, occur throughout Unit 2 where beds average ∼ 4 to 6 cm thick and are restricted to the lower part of Unit 3 where beds average ∼ 10 cm thick. Type C burrows are either partly or completely filled with framboidal pyrite and sub-equal amounts of these pyrite framboids are overgrown by poorly crystalline pyrite. Framboidal pyrite precipitated first as a byproduct of microbial degradation of labile organic matter (EPS) that lined the burrow walls and created a local micro-reducing environment with a sharp redox gradient with the surrounding oxic sediment (Konhauser and Gingras 2007; Gingras et al. 2014). The dimensional similarity of the pyrite framboids suggests that they began to precipitate at about the same time and their size (∼ 5 to 15 μm) indicates possible growth in as little as one to several weeks after burrow formation (Wilkin and Barnes 1997; Rickard 2019). Moreover, these burrows most probably remained isolated from the burrows of other benthic meiofaunal burrowers that otherwise would have introduced oxygenated water and re-established oxic conditions. These locally developed micro-reducing conditions then continued to burial depths where microbial sulphate reduction was established (Coleman 1985; Morad 1998) and facilitated pyrite precipitation (Raiswell 1982; Berner 1985). The presence of sub-micron-size crystals of iron sulphide along mica cleavage planes is indicative of in situ precipitation of iron sulphide in porewaters enriched in hydrogen sulphide, which most probably occurred within the now well-established sulphate-reduction zone, and also where requisite ferrous iron was sourced directly from the detrital mica grain (Schieber 2011b). The intense bioturbation and occurrence of framboidal pyrite overgrown by poorly crystalline pyrite throughout Unit 2 indicates that meiofaunal activity commonly influenced shallow porewater chemistry under annual sedimentation rates of ∼ 4 to 6 cm yr−1. Despite the abrupt increase in annual sedimentation rate (∼ 10 cm yr−1) that marks the onset of Unit 3, infaunal meiofauna remained resilient and were still able to inhabit the seafloor and generate mucous-lined burrows that soon-after became susceptible to diagenetic mineral growth. However, with the maintenance of these high sedimentation rates, and in combination with other paleoenvironmental controls (see next), meiofauna eventually became overwhelmed, resulting in the loss of bioturbation, locally accumulated organic matter (EPS), and associated early diagenetic iron sulphide of framboidal and poorly crystalline forms.

Paleoenvironmental Controls on Benthic Meiofauna in the Champlain Sea

At the base of the sedimentary succession (Fig. 3), a few-meters-thick interval of unbioturbated, laminated mud rhythmites (Unit 1) transitions abruptly upward into intensely bioturbated mud (Unit 2) representing the rapid change from glaciolacustrine to glaciomarine conditions (Al-Mufti et al. 2022) and coincides with the rapid incursion of Atlantic Ocean seawater into the Champlain Sea basin (e.g., Fulton et al. 1987; Parent and Occhietti 1988). The onset of bioturbation in the form of type A, B, and C burrows in Unit 2 indicates that marine conditions provided an ecologically favorable environment for burrowing meiofauna with a notable scarcity to absence of burrowing macrofauna. Burrowing intensity changes little stratigraphically upward through Unit 2 but decreases markedly at the base of Unit 3, with the abrupt appearance of diffusely banded mud. Notably, and irrespective of the marked increase in bed thickness across the Unit 2 to Unit 3 boundary, ∼ 4 to 6 cm average thickness and 15 cm maximum compared to ∼ 10 cm on average and 28 cm maximum, respectively (Fig. 4), type A, B, and C burrows remain common. This suggests that despite the increase in sedimentation, benthic meiofauna remained resilient and continued, at least initially, to exhibit fodinichnia-domichnia ethologies at the seafloor. Since each bed represents deposition during an annual meltwater discharge cycle, bed thickness, neglecting the effects of mechanical compaction, which because of minimal depth of burial (< 66 m) would have been negligible, is a measure of annual sedimentation rate. In the upper part of Unit 3 (∼ 15 meters above the base of Unit 3) average bed thickness increases abruptly to ∼ 14 cm (Fig. 4) indicating another rapid increase in sedimentation rate. Associated with the increased rate of sedimentation is a notable absence of bioturbation, although the exact stratigraphic position where bioturbation ceases, or whether the change is gradual or abrupt, is unknown due to insufficiently closely spaced samples. Nevertheless, the inverse relationship between bioturbation intensity and bed thickness suggests that sustained sedimentation rates of the order of ∼ 14 cm yr−1 were at least partly responsible for the initial reduction and then elimination of the tracemaking endobenthic fauna (e.g., Gingras et al. 2011; Boulesteix et al. 2020).

Bed thickness, and accordingly annual sedimentation rate, continue to increase from the upper part of Unit 3 (∼ 14 cm average, 20.5 cm maximum) into Unit 4a (∼ 19.6 cm average, 40 cm maximum) (Fig. 4), and burrows remain notably absent. The continued up-section absence of bioturbation may have been, in conjunction with increased sedimentation rate, partly due to the abrupt onset of hyperpycnal flows in Unit 4a (Al-Mufti et al. 2022). More specifically, freshwater-charged hyperpycnal flows associated with seasonally fluctuating glaciogenic meltwater discharge sourced from the ablating Laurentide Ice sheet may have caused salinity dilution at the sediment-water interface. Physico-chemical stresses of salinity fluctuations, increased sedimentation rate, and elevated water turbidity would have been most pronounced in the summer months when precipitation and surface runoff were at their maximum (Cowan and Powell 1991; Eyles et al. 1992), and together, probably imposed significant ecological stress on marine benthic colonization (MacEachern et al. 2005; Buatois and Mángano 2011) and may explain the scarcity to absence of any evidence of macrofaunal burrowing activity. Moreover, reduced seawater salinity was further enhanced, and most probably controlled principally by volumetrically large inflows of freshwater from glacial lakes that rapidly drained into the Champlain Sea (Teller 1988; Rodrigues and Vilks 1994; Cronin et al. 2008).

In the upper part of the succession (∼ 8 m from the ground surface), type D burrows occur, but exclusively in beds of deformed mud (Unit 4b). Deformed mud beds occur intercalated with Unit 4a strata and were interpreted by Al-Mufti et al. (2022) to be the product of episodic slope failure and downslope resedimentation along a steep deltaic margin. Prior to slope failure, mud that accumulated on the seafloor was thoroughly bioturbated by a type D benthic meiofauna. Furthermore, the exclusive occurrence of type D burrows in deformed mud blocks of Unit 4b strata, and the absence of type A, B, and C burrows in Unit 4a and 4b strata, suggests that the type D meiobenthic tracemaker and its burrowing behavior were exclusive to environmental conditions that developed in shallower, upslope areas that were more proximal to the outflow of fresh glacial waters, and therefore brackish water conditions. The change to more brackish water conditions coincided also with the shallowing of the basin in response to deltaic progradation and glacio-isostatic ground rebound (Gadd 1987; Lewis et al. 2022). Notably also, type D burrows occur in silt-rich sediment (e.g., Fig. 7A), whereas type A, B, and C occur in comparatively more clay-rich sediment, suggesting that substrate type may also have exerted an ecological stress on meiofaunal colonization.

Ice-proximal glaciomarine environments undergoing deglaciation typically generate an upward-fining succession of sandy rhythmites that grade vertically into bioturbated mud, reflecting the temporal shift from an ice-proximal to ice-distal setting and the progressive increase in distance from the sediment source (i.e., the ice sheet/glacier; Ó Cofaigh and Dowdeswell 2001). This vertical sequence of facies, namely, laminated rhythmites overlain by bioturbated mud, has been reported from several modern and ancient glaciomarine successions (e.g., Polyak and Solheim 1994; Yoon et al. 1997; Graham et al. 2010; Streuff et al. 2017), and the facies transition is commonly taken to reflect a decrease in sedimentation rate, and accordingly, an increase in faunal colonization and bioturbation (Ó Cofaigh and Dowdeswell 2001). Strata of the Champlain Sea basin show a similar succession of laminated rhythmites (Unit 1) overlain by bioturbated mud (Unit 2, lower part of Unit 3). However, in the case of Champlain Sea strata, which accumulated during progressive retreat of the continental-scale Laurentide Ice Sheet (Parent and Occhietti 1988; Cummings et al. 2011), the abrupt facies change from laminated mud rhythmites to bioturbated mud is a consequence of a change from glaciolacustrine to glaciomarine conditions and a mechanistic change in the style of sediment transport (Al-Mufti et al. 2022). This change is also marked by an abrupt increase in sedimentation rate, rather than a decrease in sedimentation rate related to the continued wasting and backstepping of the ice margin and the shift from ice-proximal to ice-distal depositional conditions. Except for the resedimented slope-derived deformed mud blocks (Unit 4b), bioturbation becomes absent in the upper part of Unit 3 and remains so up-section into Unit 4a and to the ground surface. Moreover, the elimination of meiofaunal bioturbation coincides with a critical sedimentation rate of ≤ ∼ 14 cm yr−1, implying that annual sediment accumulation rates remained high during the continued rapid ablation of the ice sheet, and significantly, led to the cessation of meiofaunal colonization at the borehole sites. The characterization of these sedimentation rates as being relatively ‘high’ is perhaps not unreasonable when compared to the sedimentation rates of Alaskan fjords in contact with glaciers, where rates of 440 cm yr−1 (Glacier Bay fjord; Powell 1981), 48 cm yr−1 (3.5 km down fjord from the Hubbard Glacier terminus; Cowan et al. 1997), and even an equivalent rate of 14 cm yr−1 (15 km down fjord from the Hubbard Glacier terminus; Cowan et al. 1997) and (6 km down fjord from the Muir Glacier terminus; Cowan et al. 1999) serve to enhance the preservation of primary physical sedimentary structures, like planar lamination, by inhibiting bioturbation.

Late Pleistocene to Holocene mud-rich (silt-clay) strata in the Champlain Sea basin exhibit characteristics indicating individual beds represent deposition related to regular fluctuations in annual glaciogenic discharge, and therein provide a highly resolved chronometer for assessing sedimentological properties like rates of sedimentation. These short-term variations are superimposed on longer-term variations in sedimentation that define four lithostratigraphic units (units 1, 2, 3, and 4a) with each unit boundary marked by an abrupt and persistent increase in bed thickness, and accordingly, annual sedimentation rate. A fifth unit, Unit 4b, occurs as interbeds in Unit 4a strata and represents sediment remobilized downslope from a shallower part of the basin. At the base of the succession unbioturbated, ∼ 2 cm-thick beds of Unit 1 are overlain abruptly by intensely bioturbated, ∼ 4 to 6 cm-thick beds of Unit 2. Bioturbation is in the form of < 1 mm-wide meiofaunal traces with a notable scarcity to absence of macrofauna. This abrupt onset of biogenic activity is associated with the sudden inflow of Atlantic Ocean water and the conversion of a glaciogenic lake (Unit 1) into the glaciomarine Champlain Sea (units 2, 3, 4a, and 4b). Meiofaunal traces comprise four types and three are differentiated by differences in the texture and fabric between the burrow fill and burrow lining or surrounding matrix (types A, B, and D). A fourth burrow type (type C) is filled with early diagenetic iron sulphide minerals that indicate bacterially mediated organic matter (EPS) degradation under sulphate reducing conditions. Type A, B and C burrows occur throughout Unit 2 and the basal part of Unit 3 (average bed thickness ∼ 10 cm), but then are absent in the upper part of Unit 3 (average bed thickness ∼ 14 cm) and Unit 4a (average bed thickness ∼ 20 cm). The loss of bioturbation beginning in the upper part of Unit 3 coincides with average annual sedimentation rates of ∼ 14 cm yr−1 and greater, suggesting this may represent a threshold rate of sediment accumulation for infaunal activity by meiofauna. Type D burrows occur exclusively near the top of the succession in slump blocks of Unit 4b, which combined with the absence of type A, B, and C burrows, illustrates the effects of physico-chemical stresses of salinity and sediment substrate type on tracemaker and/or trace-making behavior. In stratigraphic intervals containing meiofaunal burrows and iron sulphide cement, their spatial and temporal distribution provides important insight into the influence of biogenic activity on the geochemical pathways and processes in the shallow subsurface of subaqueous glaciomarine sedimentary basins. Importantly, the identification of meiofaunal traces and associated cement-filled burrows using high-resolution petrographic analysis should be further explored in future studies of other glaciomarine successions to better calibrate the critical range of sedimentation rate and other physico-chemical stresses that influence meiofaunal activity.

The authors gratefully acknowledge Drs. Sӧren Jensen, Gabriela Mángano, Maximiliano Paz, and Joonas Virtasalo for their constructive and helpful comments that helped to improve the paper. We also thank Dr. Randy Enkin for reviewing an earlier version of the manuscript. Financial support for this project was provided by the City of Ottawa, Rideau Valley Conservation Authority, and both the Groundwater Geoscience and Public Safety Geoscience programs at the Geological Survey of Canada (Natural Resources Canada). Additional support was provided by a Husky Energy (Cenovus) Fellowship in Sedimentology and Petroleum Geology to Al-Mufti and a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant to Arnott. We thank Stephen Wood (London, ON) for thin section preparation and Pasan Herath (Ottawa, ON) for help with compiling the bed thickness data. This article represents Natural Resources Canada contribution number 2230239.

Data are available from the PALAIOS Data Archive:

Open Access CC-BY 4.0

Supplementary data