Ichnological analysis in the studied section reveals the exclusive presence of Phycosiphon incertum. Trace fossils are characterized by small size, more or less cylindrical cores, and in some cases flattened, U-shaped lobes, as well as “fish-hook” shapes and pairs of black spots. Dark tubes consisting of fine-grained material are surrounded by a lighter mantle. Spreite structures were not observed, but this could be an effect of cutting. According to size, two groups of traces can be distinguished: (a) the small ones have observed tubes of a maximum length of 2–3 mm, 0.4–0.5 mm width, and marginal tubes approximately 0.2 mm wide; (b) the large forms have tubes 5–10 mm long, 0.8–1 mm wide, and marginal tubes 0.4 mm wide (Fig. 3). Small forms are dominant in the studied sections. The presented features allow to assign all the discussed trace fossils to P. incertum Fischer-Ooster.

According to the emended diagnosis by Wetzel & Bromley (1994), P. incertum refers to an extensive, small-scale spreite trace fossil consisting of repeated narrow, U-shaped lobes enclosing millimetre- to centimetre-scale spreite, branching regularly or irregularly from an axial spreite of similar width. Lobes are protrusive and mainly parallel to bedding, but the plane enclosing their width may lie horizontally, obliquely or even vertically with respect to the bedding plane. This is the widely accepted ichnotaxonomy by synonymization of Anconichnus horizontalis with P. incertum, emending the original identification of P. incertum by including non-bedding-parallel specimens, and showing Anconichnus as a junior synonym of Phycosiphon. A clear description of P. incertum is presented in Bromley (1996: 264–266). The author refers to a complexly-lobed, small-scale spreite structure, surrounded by a thin mantle of pale sediment; the core consists of backfilled dark material, and the spreite is made of the same pale material as the mantle. When mantle and spreite are difficult to detect, as in a pale matrix, the paired sections of fill where lobes are cut transversely, or “fish-hooks” where they are cut longitudinally, permit conclusive identification. Illustrative reconstructions of the multiple phases of foraging can be found, including three-dimensional models of “phycosiphoniform” burrows (i.e., figure 11.11 in Bromley 1996; Bednarz & McIlroy 2009). See Uchman (1995, 1998 1999) for discussion of the Phycosiphon group, and ichnotaxonomy at the ichnospecies level.

Detailed analysis of P. incertum throughout the studied section reveals several patterns according to distribution, abundance and density, usually related to the differentiated size-groups: (a) sparsely and randomly distributed larger forms consisting of disperse, isolated P. incertum in a weakly to non-laminated intervals, with variable orientation in respect to the bedding plane, (b) sporadic records of small forms within a laminated interval, occurring as isolated specimens or as patches of several burrows with variable orientation, (c) densely packed small P. incertum, mainly parallel to the bedding plane, being located as horizons in laminated intervals and (d) densely distributed small forms (mainly horizontally) occupying a fully bioturbated interval several centimetres in thickness.

Variations in P. incertum features within the laminated to bioturbated intervals include: (a) fully laminated intervals, consisting of millimetre-scale laminae with scarce or even absent P. incertum, (b) highly bioturbated intervals, either sparsely bioturbated with larger P. incertum, or intensely bioturbated intervals with small-sized forms and (c) bioturbated to laminated intervals, appearing as weakly laminated intervals with densely packed horizons or as patches of small P. incertum.

An idealized stratigraphic pattern of the differentiated cases is as follows: (a) at the base, a laminated interval with absent or only very scarce P. incertum; (b) above, a coarse-grained interval with cross-bedding lamination, showing a slightly undulated (erosional?) at the base and a gradual transition with a bioturbated interval at the top; and (c) finally, a gradual transition to a laminated interval with absent or very few P. incertum. This idealized pattern, however, may be present in a number of variations related to the presence of bioturbated horizons in the laminated/non-bioturbated interval (see below).

The Gilsonryggen Member of the Frysjaodden Formation consists of prodelta sediments with distal turbidites. In the studied section, several cases were distinguished (Figs. 4, 5, 8).

Case A is present in the transition between successive turbiditic sequences: the transition shows a small erosional undulate-shape body composed of the massive—centimetre-thick—graded Ta interval of the Bouma sequence. It cuts into a laminated mud interpreted as the Te interval of the Bouma sequence below. This massive graded interval shows comparatively coarse grains, is lighter in colour, consisting of discrete, scarce, larger P. incertum (Fig. 4a). The laminated interval is characterized by the absence or scarce record of both, trace fossils and foraminifera.

Case B is observed in the middle part of a turbiditic sequence. A centimetre-thick cross-laminated interval Tc of the Bouma sequence is registered above the lower laminated interval Tb, with a slightly erosional undulated surface in between. The cross-laminated interval Tc is characterized by the absence/scarce record of P. incertum and foraminifera, as occurs in the former lower laminated interval (Fig. 4b, c).

Case C (middle part of a turbiditic sequence) is characterized by a millimetre-thick cross-laminated interval (Tc) above the lower laminated interval (Tb), with a slightly erosional undulating surface in between. This millimetre-thick cross-laminated interval also shows the absence or only scarce occurrence of P. incertum. This case is similar to case B, but associated to a thinner cross-laminated interval (Fig. 4d). In some cases, a cross-laminated interval is not observed; the bioturbated interval (Td) therefore overlies the laminated interval (Tb).

Case D refers to the middle–upper parts of the turbiditic sequence. A gradual transition from bioturbated to laminated interval (Td–Te), showing progressively decreasing size and abundance of P. incertum and disappearance of the (scarce) foraminifera (Fig. 5a) is observed here.

Case E (middle–upper parts of the turbiditic sequence) shows a gradual transition from a bioturbated interval with disperse, abundant, large P. incertum, to a thick laminated interval showing alternating horizons with small Phycosiphon and thin laminated horizons (Td–Te) without bioturbational structures (Fig. 5b).

Case F (middle–upper parts of the turbiditic sequence) displays a gradual transition from a bioturbated interval with disperse, abundant, large P. incertum, to a thick laminated interval showing patches of small Phycosiphon and ending in a laminated horizon without bioturbational structures (Td–Te; Fig. 5c).

Geochemical analyses of the studied section show a generalized pattern for all the studied proxies—the presence of long-term trends and the absence of short-term fluctuations, together with the presence of scattered, significant peaks of increasing geochemical values. The record of Phycosiphon is not directly related to increases in palaeoproductivity proxies, yet the peaks of redox proxies always occur in laminated horizons.

From the studied section, interval 1 (222.66–222.30 m; 34 cm thick) was selected to disclose the observed pattern (Fig. 6). All palaeoproductivity proxies (P/Ti, Ba/Al, Sr/Al and U/Al) show maximum values and a peak coinciding with horizons poorly-bioturbated by Phycosiphon. All redox proxies (Co/Al, Cu/Al, Ni/Al, Mo/Al and Mo_aut) show no variations along the studied interval, except for an important peak, with maximum values, just above the peak of the palaeoproductivity proxies, related to the well-laminated interval directly above the poorly-bioturbated horizons.

The analyses with back-scattered electron imagery reveal further differences between the well-laminated intervals and bioturbated intervals. The well-laminated intervals are richer in clay minerals, pyrite framboids and phytodetritus than the bioturbated intervals (Fig. 7). The size of the pyrite framboids is lower in well-laminated intervals (3.6–5.5 µm) than in bioturbated intervals (6.4–11.5 µm).

P. incertum traces are interpreted as having been produced by deposit feeders, usually opportunistic, at different tiers in mud and fine sand at variable depths up to 15 cm in sediments deposited in water depths below lower shoreface (Fu 1991; Goldring et al. 1991; Wetzel & Bromley 1994; Uchman 1995, 1999; Wetzel & Uchman 2001; Hovikoski et al. 2008; Wetzel 2010). Phycosiphon is considered a pascichnial/fodinichnial structure (Ekdale & Masson 1988; Bromley 1996); the absence of spreite was interpreted as reflecting a pascichnia (grazing) rather than a fodinichnia (mining) behaviour (Ekdale & Masson 1988). Phycosiphon traces have been observed in recent muddy mass-flow deposits in deep-marine settings, but the producer has not yet been recognized (Wetzel 2008). According to Bednarz & McIlroy (2009), producers of phycosiphoniform burrows were small, probably vermiform organisms.

Phycosiphon is frequently associated with food availability and oxygen conditions. The Phycosiphon producer reworked a relatively small volume of the sediment containing abundant nutritious material. Depth–size grading is typical of colonization; after initial colonization, the trace-maker grew in size and burrowed deeper into the sediment (Wetzel & Uchman 2001; Hovikoski et al. 2008). Orientation of Phycosiphon spreiten varies from inclined in homogeneous muds to horizontal in laminated deposits, probably reflecting the homogeneous distribution of food material (Wetzel & Uchman 2001; Wetzel 2010). Phycosiphon is common in poorly oxygenated sediments, but the absence of a connection to the seafloor indicates that immediate colonization took place in an oxygenated upper layer habitat (Ekdale & Masson 1988; Wetzel & Uchman 2001).

Integrated lithological and foraminiferal analyses of the core shed light on the general palaeoenvironmental conditions (Nagy et al. 2013). The upper part of the Gilsonryggen Member (Frysjaodden Formation) is related to the initial phases of delta progradation concomitant with the development of a highstand system tract. Salinity is a main restricting factor, though it decreases upward (Nagy et al. 2013), but short-term fluctuations may also be considered. The dominance of opportunistic organisms, such as that producing P. incertum, must be interpreted in the context of a stressed habitat resulting from salinity changes and fluctuations. However, the observed ichnological variations suggest the influence of additional environmental factors, including oxygenation and benthic food availability. The scarceness of foraminifera and the restricted occurrence of oxygen-tolerant forms (Thurammina papillata, Psammosphaera fusca and Trochammina inornata [Nagy et al. 2013]) would confirm the low oxygen availability.

The exclusiveness of P. incertum, the well-developed laminated intervals and the overall dark colour of the deposits all point to anoxic or highly dysoxic sediments. A minor increase in oxygen content to lower dysoxic conditions allowed for a localized colonization by the producer of Phycosiphon. The offset between the peak of palaeoproductivity proxies and the redox proxies may also be related to a consumption of oxygen resulting from increased productivity. Anoxic conditions are corroborated by the small size of pyrite framboids. Pyrite framboids only form at the redox boundary where oxygen-bearing and hydrogen sulphide-bearing waters are in contact (Wilkin et al. 1996; Reolid 2014). In oxygenated bottom water settings, pyrite framboid growth occurs in the upper sediment column, just below the sediment—water interface. However, in euxinic settings the redox boundary is within the water column. Framboids growing in the water column have a limited diameter range, as once they have grown to a certain size, they will sink out of the narrow redox interface zone and into oxygen-free bottom water and sediment where further growth is not possible (Wilkin et al. 1996; Dustira et al. 2013). The smallest mean diameters (3–5 µm) with a very limited size range are indicative of euxinic conditions (Bond & Wignall 2010). Framboids forming within the sediment under oxic bottom water conditions attain greater diameters on average (7.7 µm) as they reside and grow at the redox interface zone for a longer time (Wilkin et al. 1996). According to Wilkin et al. (1996) and Dustira et al. (2013), mean framboid size <7.7 µm indicates dysoxic conditions, between 5 and 7.7 µm corresponds to dysoxic—anoxic conditions and <5 µm indicates euxinic conditions. Taking into account these ranges of pyrite framboid size, the laminated intervals of study may have developed under euxinic conditions because the mean size observed was 3.6 µm. Since the mean size in the bioturbated intervals was 8.3 µm, framboids most likely developed in the sediment pore-water there.

Within deltaic systems, Phycosiphon has been observed mainly in sediments of the delta front, the transition zone between delta front and prodelta, and in the prodelta (i.e., Bann et al. 2008; Buatois et al. 2008; Carmona et al. 2009; Buatois et al. 2011; Weiguo et al. 2011; Buatois et al. 2012). Several papers addressed the immediate colonization of event beds by Phycosiphon trace-makers (Stow & Wetzel 1990; Goldring et al. 1991; Wetzel & Balson 1992; de Gibert & Martinell 1998). In muddy turbidites Phycosiphon producers are interpreted to be the first organisms colonizing the upper part of a turbidite bed, probably very rapidly, only a short time after deposition (Wetzel & Uchman 2001).

During delta progradation at the studied sections, turbidite deposition was extended, showing a succession of Bouma sequences. These Bouma sequences graded from massive sand deposits at the base to laminated silt and mud, with hemipelagic and pelagic material on top. Models of tiering, trace fossil distribution and ichnofabrics for turbidite/couplets have been proposed (Uchman 1999; Uchman & Wetzel 2011, 2012). Often, a sequential colonization of event layers can be found (Wetzel & Uchman 2001). Newly deposited turbiditic sediments contain relatively well-oxygenated pore water and a relatively high amount of benthic food, allowing initial colonization by small opportunistic deposit feeders, such as mining Phycosiphon producers. Later, when oxygen and food decrease at the level where mining as feeding strategy becomes inefficient, the sediment is colonized by stationary chemosymbionts, such as Chondrites producers. Finally, during the long period before deposition of a new turbidite, the slowly deposited background sediment is colonized by graphoglyptids in well-oxygenated settings (Uchman & Wetzel 2011, 2012). The studied case reveals clear differences with the model proposed by Uchman & Wetzel (2011, 2012), probably related to less favourable conditions, impeding a more diverse and better developed trace fossil assemblage.

When interpreting the turbiditic sediments in the studied section, the particular environmental conditions have to be taken into account, especially a long-term trend of decreasing salinity with short-term fluctuations, an oxygen content fluctuating between anaerobic and dysaerobic conditions, a short-term local input of benthic food and episodic sedimentation affecting all these factors. Variations in the ecological and sedimentological parameters would explain the variable record of the encountered Bouma sequences (Fig. 8) (e.g., Mulder 2011). Their variable development is interpreted according to the observed sedimentological and ichnological features associated with the transition between successive turbiditic sequences or within a single turbidite sequence (Fig. 8). We put forth six cases, as follows.

Case A: input of coarse sediment (Ta) was associated with an overall increase in oxygenation from the anaerobic laminated interval (Te) below to dysaerobic, conditions in Ta, probably related to a relatively high amount of food, allowing a disperse colonization of larger Phycosiphon trace-makers.

Case B: deposition of the cross-laminated interval was not associated with an increase in oxygen content sufficient to allow a colonization by Phycosiphon trace-makers; anaerobic conditions during deposition of the upper part of the previous laminated interval (Tb) continued until the base of the next bioturbated interval (Td) above the cross-laminated interval was deposited.

Case C: this case, similar to case B, is characterized by a thin cross-laminated interval that could be related to rapid deposition of the cross-laminated interval or to minor erosion. Absent or only scarce P. incertum can be interpreted as reflecting persistent low-oxygen conditions (as in case B) or as evidence of a short time of deposition for the thin cross-laminated interval before deposition of the next interval (Td) which impeded colonization by Phycosiphon trace-makers.

Case D: this case could be associated with a gradual decrease in oxygen content from dysaerobic to finally nearly anaerobic conditions, documented by the colonization by large Phycosiphon trace-makers, then by small Phycosiphon and finally lacking bioturbation.

Case E: this case reflects a gradual decrease in oxygenation from dysaerobic conditions, allowing a colonization by large Phycosiphon trace-makers, to nearly anaerobic conditions generally preventing burrowing activities which are, registered only during short-time improvements in oxygenation.

Case F: in this case, the oxygen content decreased from dysaerobic conditions, allowing a colonization by large Phycosiphon trace-makers, to nearly anaerobic sediments strongly reducing the activity of Phycosiphon trace-makers. However, patches of Phycosiphon could be associated with short-time improvements in oxygenation together with localized patches of available benthic food, probably related to minor flows that brought oxygen and food from shallower zones.