INTRODUCTION
Deep-sea fan systems are significant reservoirs of buried organic carbon, making them an important part of the global carbon cycle (e.g., Baudin et al., 2010; Cunningham & Arnott, 2023; Galy et al., 2007; Hage et al., 2022; Hussain et al., 2021; Masiello, 2007).
Organic carbon has been reported from a range of deep-marine sedimentary environments, including basin-floor lobes and slope channels and levees (Baudin et al., 2010, 2017, 2020; Cunningham et al., 2023; Cunningham & Arnott, 2023; Hussain et al., 2021; McArthur et al., 2017; McArthur, Kneller, Souza, et al., 2016; McArthur, Kneller, Wakefield, et al., 2016; Saller et al., 2006; Stetten et al., 2015), and that turbidity currents have been shown to be effective mechanisms for its transport from shallow continental margins to more distal parts of the depositional system (e.g., Saller et al., 2006; de Baudin et al., 2010, 2017, 2020; Cunningham & Arnott, 2023; Galy et al., 2007; Hage et al., 2020, 2022; Hussain et al., 2021; Lee et al., 2019; Masiello, 2007; Stetten et al., 2015). Studying the occurrence and stratigraphic distribution of organic-rich strata can provide insight into the physical processes that govern the transport and deposition of organic matter to the deep ocean and source-to-sink trends from the continental realm to the deep marine.
Recent work on Neoproterozoic slope deposits in Western Canada has shown that organic carbon occurs mostly in thick-bedded sandstones in levee deposits (Cunningham et al., 2023; Cunningham & Arnott, 2023). Although the physical and chemical processes that occur during the deposition of a these organic-rich beds have previously been described in detail by Cunningham and Arnott (2023) and shown to be related to various external environmental factors, such as primary productivity, sea level, weathering, and sediment flux (Cunningham et al., 2023), the characteristics of the organic-rich flows themselves have received much less attention. Here, we present descriptions of the stacking patterns of organic-rich beds from this site and discuss the size and frequency of flows that transport organic matter to the deep-marine, in addition to their possible triggering mechanisms. Importantly, these results help to identify basin-scale factors that control the distribution of organic carbon in deep-sea sediments.
GEOLOGICAL SETTING
The study area is located in the Cariboo Mountains of east-central British Columbia, Canada (Figure 1) where strata of the Neoproterozoic Windermere Supergroup crop out extensively and record the breakup of Rodinia and subsequent sedimentation along a passive continental margin. The base of the succession consists of intercalated glacial diamictites and mafic volcanics related to early rifting (Eyster et al., 2018). This is overlain by a post-rift succession of deep-water (Kaza Group and Isaac Formation) to shallow shelf and platform sedimentary rocks (Cunningham and Yankee Belle formations) (Campbell et al., 1973) that record the progradation of the passive continental margin into the thermally subsiding proto-Pacific Ocean (Figure 2) (Ross, 1991; Ross et al., 1995; Ross & Arnott, 2007).
Age control of the Windermere Supergroup in this area is generally poor due to a paucity of datable markers and absence of biostratigraphy; age being largely constrained by radiometric dates from below (736 – 728 Ma; Evenchick et al., 1984; Eyster et al., 2018; McDonough & Parrish, 1991) and above (570 +/- 5.3 Ma; Colpron et al., 2002) its bounding unconformities. Within the Windermere succession dates are limited to a maximum depositional age of 650 Ma based on U-Pb dating of detrital zircons (Hadlari et al., 2021), and a Re-Os isochron date of 607.8 +/- 4.7 Ma from organic-rich mudstones of the Old Fort Point Formation (Kendall et al., 2004; Smith et al., 2014).
Study Area
The Castle Creek study area is located on a vertically dipping limb of a southwest-verging anticline. Strata are recently deglaciated, vegetation-free, and superbly exposed in a section approximately 2.6 km thick and up to 8 km wide (Navarro & Arnott, 2020; Terlaky et al., 2015). Basin-floor deposits of the upper Kaza Group make up the lower ~800 m of the exposed succession and are overlain by ~1.8 km of slope deposits of the Isaac Formation (Figure 2). The Isaac Formation comprises at least seven up to 200 m-thick, coarse-grained, laterally discontinuous sandstone and conglomerate units, surrounded by thin-bedded, mudstone-rich strata. These discontinuous sandstones and conglomerates are interpreted to be the fill of slope-channel complexes bounded by finer-grained, genetically related levees (Figure 3), and are informally termed Isaac Channel Complex 0 – 6 (ICC0 – ICC6). The levee deposits of ICC3 and ICC4 are particularly well-developed, and their sedimentology, stratigraphic architecture, and geochemistry have been described in detail by Khan and Arnott (2011), Khan et al. (2011), Cunningham and Arnott (2021, 2023); Bergen et al. (2022), and Cunningham et al. (2023). Notably, although organic-rich strata are generally uncommon in the study area, organic-rich beds are abundant in a 60-m-thick interval in levee deposits below ICC4 (Cunningham et al., 2023; Cunningham & Arnott, 2023).
METHODS
Levee deposits at the Castle Creek study site were logged in bed-by-bed detail to capture trends in lithology, stratal thickness, sedimentary structures, and grain size. This was combined with geochemical analyses, principally total organic carbon (TOC) (Cunningham et al., 2023; Cunningham & Arnott, 2023), to identify organic-rich strata and its stratigraphic distribution. In total 340 m of continuous vertical stratigraphy was logged in cm-scale detail (Figure 3). Stratigraphic logs were spaced 40-70 m apart along the strike of the outcrop and individual beds were correlated over ~ 2 km in the field and later combined into a single composite log. Turbidites are described using the five-division terminology of Bouma (1962).
Turbidites rarely contain all five divisions; instead typically comprising only part of the full sequence. Lower-division turbidites consist of the bottom two divisions (Ta and Tb), whereas upper-division turbidites comprise only the upper two or three divisions (Tc, Td, and Te).
RESULTS
Organic-rich strata at the Castle Creek study site are primarily confined to one 60-meter-thick stratigraphic interval in levee deposits immediately below ICC4, which previously was reported to correlate with conditions of elevated sea level and primary productivity on the continental shelf (Figure 4) (Cunningham et al., 2023; Cunningham & Arnott, 2023). Organic-rich beds account for ~5% of the measured stratigraphy in this study and are divided into two facies: organic-rich mudstone and organic-rich sandstone turbidites. TOC in these organic-rich facies ranges from 0.4 – 4.04 % (Cunningham & Arnott, 2023), which if corrected for lower greenschist metamorphism (and the loss of ~50-80% of the original organic content) may be of the order of 2 –16% (Hayes et al., 1983; Smith et al., 2014; Tissot & Welte, 1978). Organic-rich sandstone turbidites account for ~ 85% of the organic-rich strata at this site. The sedimentology and geochemistry of these facies are summarized below; for detailed descriptions see in Cunningham and Arnott (2023).
Levee deposits at the Castle Creek study site are dominated by very thin- to thin-bedded (1 – 15 cm), upper-division turbidites (~72% of turbidites in the measured section) (Bergen et al., 2022; Cunningham & Arnott, 2021; Z. A. Khan & Arnott, 2011). Interspersed with these levee thin beds are uncommon, organic-poor, medium- to thick-bedded (generally 20 – 80 cm thick), lower-division turbidites and organic-rich sandstone turbidites. These latter strata range from 30 – 140 cm thick and are distinctively banded, with alternating layers of orange, highly dolomite-cemented sandstone, and black, mud-rich sandstone (Figure 4). Most of the organic matter in these beds occurs in the black bands as micro- to nano-scale coatings on clay particles, although uncommon sand-sized amorphous grains and organo-mineralic aggregates are also present in the cemented orange bands. The unique banding in these beds (informally termed “tiger-striped” beds) is interpreted to result from alternating periods of turbulent and cohesive flow conditions as organic-rich turbidity currents overspilled channel margins and deposited on the levee (Cunningham & Arnott, 2023). The high cement content (up to 35%) in the orange bands, which consists of ferroan dolomite, is interpreted to have formed from processes related to the diagenesis of the organic matter (Boles, 1978; Cunningham & Arnott, 2023; Curtis, 1978; Kelts & McKenzie, 1982; Mazzullo, 2000).
Organic-rich sandstone turbidites occur as single beds that are separated by 2 – 22 m in a thick succession of comparatively organic-poor (< 0.4% TOC) thin-bedded turbidites. These beds are laterally continuous over distances of several hundreds of meters and show only a slight thinning away from the channel. Although sand-rich tiger-striped beds occur exclusively in the single 60-m-thick interval below ICC4, rare organic-rich mudstone beds crop out elsewhere in the 340-m thick succession (Cunningham et al., 2023). Deformed organic-rich mudstones have also been observed in mass transport deposits (slides and debrites) at Castle Creek, which notably also contain fragments of oolitic and stromatolitic limestone (Bergen et al., 2022).
DISCUSSION
The concentration of organic matter in anomalously thick, sand-rich turbidites provides important information about both the source and depositional history of these beds. Previous work on levee deposits, both at Castle Creek and elsewhere, has shown that thick, coarse-grained, lower-division turbidites are generally the result of large, sand-rich turbidity currents whose thickness is significantly greater than the height of the channel, resulting in continuous overspill of coarse-grained sediment onto the levee, or as part of overbank splay deposits (Arnott, 2010; Cunningham & Arnott, 2021; Kane et al., 2007; Peakall et al., 2000; 2023). This contrasts with the conditions for thin-bedded levee turbidites, where the lower, coarser-grained portion of the flow remains mostly confined in the channel and only lesser amounts of fine-grained sediment from the upper, dilute portion of the flow overspill onto the levee (e.g., Bergen et al., 2022). Since most of the organic carbon at the Castle Creek study area occurs in unusually thick, sand-rich turbidites, this suggests that organic matter was primarily delivered to the deep marine by anomalously large, coarse-grained turbidity currents, which were of sufficient thickness to continuously overspill the channel margins. Although it is possible that the organic-rich sandstone beds are overbank splay deposits formed when the lower, coarser-grained portion of channelized turbidity currents escape confinement or breach the levee, they differ markedly from other overbank splay deposits at the site, making this unlikely. Splay deposits typically consist of erosionally based, thick-bedded, coarse-grained, structureless and matrix-rich sandstone beds with distinctive along-strike facies changes described in earlier works (e.g., Angus et al., 2019; Bergen et al., 2022; Z. A. Khan & Arnott, 2011; Ningthoujam et al., 2022; Terlaky & Arnott, 2014). Organic-rich sandstone beds contain only fine sand, show only minor erosion at their base, and lack any significant along-strike facies change supporting the interpretation that they derive from the finer-grained, upper part of turbidity currents as they overtopped channel confinement.
Furthermore, because strata are Neoproterozoic in age, the preserved organic matter is exclusively marine and would have originated in warm, shallow waters on the continental shelf and over the upper slope (Behrensmeyer et al., 1992; Butterfield, 2014; Cunningham et al., 2023; Cunningham & Arnott, 2023; Gensel, 2021). The anomalously large flows that mobilized and transported this organic matter downslope would therefore probably also have originated on the shelf rather than the continental slope, where, in contrast, organic matter would be in low abundance or absent. Similarly, interspersed mass-transport deposits with abundant organic matter and scattered oolite and stromatolite fragments are likely to also have originated much higher on the slope, possibly near the shelf-slope break (Bergen et al., 2022). In comparison, the thick-bedded, sand-rich turbidites that are organic-poor are interpreted to reflect periods of low primary productivity on the shelf, or to have originated from slope-derived turbidity currents rather than the shelf.
There are several possible mechanisms for triggering unusually large, shelf-derived flows. For example, earthquakes are known to initiate exceptionally large turbidity currents that flush canyon systems and transport hundreds of cubic kilometers of sediment into the deep sea (e.g., Hsu et al., 2008; Krause et al., 1970; Masson et al., 2011; Mountjoy et al., 2018; Piper & Normark, 2009). Alternatively, recent work monitoring modern continental shelves and canyon systems, such as the Congo River system and the St. Lawrence Estuary, have shown that local shelf conditions can also be conducive to the formation of large flows (e.g., Hage et al., 2019; Normandeau et al., 2020; Talling, 2014; Talling et al., 2022). For example, Talling et al. (2022) recorded the longest runout turbidity currents ever measured in the Congo submarine canyon and linked them to river floods and rapid sediment accumulation at the mouth of the river with later downslope remobilization. However, this is dependent on the deep-water system being linked directly to the river system, for example the modern Congo Canyon and Laurentian Channel. Large storms, especially those with prolonged periods of sustained high winds and wave height (including but not limited to hurricanes and typhoons), have also been linked to large, long-duration turbidity currents (e.g., Normandeau et al., 2020; Porcile et al., 2020; Sequeiros et al., 2019; Zhang et al., 2018). These storms have also been theorized to contribute to river flooding and sediment delivery and accumulation in the river mouth and on the shelf, potentially priming the system for a later downslope transport event. Other potential causes of flooding that could be linked to the triggering of turbidity currents include high seasonal snowmelt or glacial outburst floods (e.g., Bornhold et al., 1994; Russell et al., 2006; Shaw & Gilbert, 1990; Tweed & Russell, 1999; Veh et al., 2019; Weckwerth et al., 2019).
Although all these mechanisms could account for the deposition of thick-bedded, sand-rich levee turbidites, this type of bed is markedly more common in the 60-m-thick organic-rich interval (two thirds of these beds occur here). In this interval, these thick beds are enriched in OM and spaced 2 – 22 m apart. This irregular stratigraphic occurrence suggests that anomalously large flows may not only be necessary to mobilize and transport significant amounts of organic matter downslope, but that elevated OM content on the shelf may be partly responsible for triggering such flows and that the frequency of these outsized turbidity currents may be linked to productivity on the outer shelf and over the upper continental slope.
In this model, high rates of organic matter production and fallout would result in the rapid buildup of OM on the seabed that would help to bind and stabilize accumulating sediment (e.g., Bhaskar & Bhosle, 2005; Winterwerp & van Kesteren, 2004; Van Leussen, 2011). These effects then resulted in an overthickening and progressive oversteepening of the sediment pile along the margins of canyons and near the shelf-slope break, which eventually becomes gravitationally unstable resulting in episodic seabed failures that generate large, surge-like flows that are considerably thicker than the depth of the slope channels through which they travel. Accordingly, the flow continuously overspills the channel margins and deposits an anomalously thick, sand- and organic-rich bed. These periodic events not only deplete the shelf of OM, but apparently reduce the slope of the local seabed and result in more typical smaller, organic-poor turbidity currents between these outsized events while the organic-rich sediment pile builds up again. Random, short-lived oceanic events such as upwellings or algal blooms may also have temporarily elevated primary productivity and contributed to the rapid accumulation of organic matter over the outer shelf and upper slope. Thick-bedded, organic-poor sandstone turbidites are interpreted to be less common than thick-bedded, organic-rich sandstone turbidites because without the stabilizing effects of cohesive OM, requisite overthickening and oversteepening of the sediment pile on the shelf and large-scale collapse occur much less frequently, and that these flows initiated from a source located on the OM-impoverished continental slope rather than the continental shelf.
It is also possible that the emplacement of thick, organic-rich turbidites could be related to the formation of large quantities of fluid mud on the shelf. Along modern coasts, particularly near river mouths or estuaries, high primary productivity and OM accumulation aid in the development of fluid mud conditions in the near-bed region by forming large, cohesive, water-saturated aggregates with clay grains (Bachmann et al., 2005; McAnally et al., 2007; Mehta, 1991; Mehta et al., 2014; Otharán et al., 2018). Being unconsolidated and water-saturated, fluid mud is generally easily mobilized (e.g., Anthony et al., 2014; Blanton et al., 1999; McAnally et al., 2007; Otharán et al., 2018; Schrottke et al., 2006; Winterwerp et al., 2017) and in turn has the ability to spawn outsized, organic-rich turbidity currents. However, we consider this interpretation unlikely, since to deposit the organic-rich sandstone beds described here, these fluid mud flows would have to self-ignite (Parker et al., 1986) and incorporate sand as they travelled downslope, as the original suspensions would probably have lacked sufficient sand content.
Significantly, these results suggest that outsized turbidity currents are the primary mechanism for organic matter delivery to the deep sea, and that flow size and frequency, in addition to primary productivity, exerts an important control on the distribution of organic carbon in deep-sea sediments. However, it is important to note that the Windermere turbidite system is Neoproterozoic in age and therefore all the OM was marine and microbial or algal in origin, which then raises the possibility that the conditions described here are unique to deep-sea turbidite systems before the evolution of metazoans or terrestrial plants. Future work could continue to examine Phanerozoic systems to compare how organic matter is sourced and preserved in similar deep-marine settings.
CONCLUSIONS
Observations from well-exposed deep-marine levee deposits of the Neoproterozoic Windermere Supergroup demonstrate that organic matter occurs primarily in thick-bedded, sandstone turbidites with distinct organic-rich and cement-rich bands. These thick, coarse-grained, organic-rich beds are interspersed among more typical organic-poor, thin-bedded levee turbidites and are interpreted to originate from anomalously thick, continental shelf-derived turbidity currents that continuously overspilled the channel and deposited sediment on the levee. It is proposed that rapid accumulation of organic matter on the continental shelf during periods of high primary productivity contributes to an overthickening and oversteepening of the sediment pile that episodically resulted in outer continental shelf instability and failure and frequent outsized flows. These outsized, continental-shelf- rather than continental-slope-initiated turbidity currents mobilize and transport vast quantities of organic matter and sediment and may be a primary mechanism for organic carbon delivery to deep-marine environments.