replying to L. Löwemark & A. Singh Communications Earth & Environment https://doi.org/10.1038/s43247-024-01461-0 (2024)

In our original contribution, we demonstrated that in the Arctic Ocean, authigenic carbonate precipitates affect foraminifera shells used for paleoceanographic reconstructions. We proved the presence of such secondary precipitates using a wide range of mineralogical, geochemical, isotopic, and optical tools. Such precipitates are not commonly seen to this extent on foraminifera in other ocean basins, however, PI_3 specimens (containing detectable amounts of overgrowth)1,2 are obviously no suitable tools to reconstruct environmental conditions during the growth of the organism. We also showed that previous reconstructions are likely to include effects of such authigenic overgrowth, also affecting radiocarbon values. In our Reply, we concur with Löwemark and Singh that bioturbation may bias Arctic records in specific instances, as outlined in the original publication. However, the specifically Arctic problem of the widespread authigenic overgrowth persists, independent of any possible bioturbation.

In their comment on our paper Löwemark and Singh address the bioturbation, an important process known to occur globally in oxygenated deep-sea sediments3,4,5,6. They identify Zoophycos in kastenlot core (KC) PS2185-6, box core (BC) PS72/413-3, and mostly (Figs. 2, 3) in gravity core (GC) PS72/413-5, which is located 1.3 km away from BC PS72/413−3 used in our study. GC PS72/413-5 (Fig. 2a Comment) has a condensed lithology, therefore, the BC does not contain this older high bioturbation unit shown in Figs. 2, 3 of the Comment. Partly intense bioturbation in the low-sedimentation records of the central Arctic Ocean has been described already in the earliest studies7. Here, bioturbation is usually most intense in the brown layers (B), assigned to interglacial or interstadial conditions8. Deep-reaching traces are most obvious in the gray layer below B37,8 (Fig. 1 of this reply and Figs. 2, 3 of the Comment). In our manuscript, we did not discriminate between different trace fossils but collectively referred to the dark brown mottles in the gray sediments below B1 as bioturbation. In these sediments, cool-white to off-white PI_2 specimens of the planktic foraminifera Neogloboquadrina pachyderma were regarded as best-preserved autochthonous specimens, whereas pristine translucent PI_1 specimens were assigned to dark brown trace fossil fillings within this gray unit (Figs. 1, 2). The coloration of our Zoophycos traces indicates that the filling consists of sediments of the B1 layer assigned to Holocene to GI 1/Bølling age8 (Fig. 1), which matches the radiocarbon ages of our PI_1 specimens1,2. The presumption of Löwemark and Singh that PI_2 specimens of MIS2-3 age originate from these trace fossils, contrasts with the presumed trace fossil age/color and the increased food demand of Zoophycos creating organisms indicating a Holocene formation6. Furthermore, it contrasts with the synchronous δ13C downcore variations in PI_2 and PI_3 specimens (Fig. 5a, e1), which we would not expect from bioturbated vs. in situ specimens. Brownish-discolored specimens PI_4 were excluded from AMS dating in Wollenburg et al.1, but are provided for comparison from the 2.5 cm-sample (Fig. 1). Since overgrowth increases with sediment depth in glacial sediments, the assumption of extremely low-sedimentation in MIS29,10 cannot be upheld until being proven by radiocarbon analyses on correspondingly diagenetically unchanged shells. We isolated PI_2 and PI_3 specimens that matched in coloration and transparency and were closest to the biogenic shell. Within the BC, the appearance of PI_2 specimens changed from translucent in the Holocene to cool-white in the deglaciation and off-white in glacial sediments. As off-white is not the original color of an N. pachyderma shell, these glacial PI_2 shells were diagenetically altered to variable degrees. Consequently, some glacial PI_2 measurements failed, generating ages comparable to PI_3 specimens, whereas, at 22.5 cm the preservation improved and even allowed for an additional high-precision radiocarbon measurement of a large sample (1000 µgC) which resulted in reasonable radiocarbon ages supporting previous measured on small samples (<100 µgC) (see Mollenhauer et al.11 for details on uncertainties of both 14C analyses). Support for the MIS2/3 radiocarbon ages of the original manuscript comes from PI_2 Cibicidoides wuellerstorfi measurements from nearby BC PS72/396, which revealed a B3/W3 border age of 27.5 C14 ka (Fig. 2).

Fig. 1: Radiocarbon ages and lithological units of sediment cores are discussed in this reply.
figure 1

a BC PS72/413-3. a1 Downcore distribution of Neogloboquadrina pachyderma PI_1−31 and exemplified discolored PI_4 radiocarbon ages. Green dot exemplified radiocarbon measurement of discolored PI_4 N. pachyderma (14.61 ka 14C). Dark pink shading = major overgrowth, light pink shade = moderate overgrowth1. a2 Carbon isotope values. a3 Core image of BC PS72/413-3, Lat. 80.277900, Long. −178.515100, water depth 1263.0 m, Recovery 0.43 m used in Wollenburg et al.1,2; as the W3 layer is disturbed by bioturbation the position in this figure follows the peak abundances of dolomite and high-magnesium calcite (Fig. 8d1). b Core image of GC PS72/413-5, Lat. 80.288800, Long. −178.483600, water depth 1274.0 m, Recovery 6.44 m used by Löwemark and Singh; note condensed B-section c Linescan of BC PS72/396−3, Lat. 80.586600, Long. −162.360700, water depth -2731.0 m, Recovery: 0.43 m. Dashed red lines indicate the boundary of the W3 layer, a sedimentary event with splendid white-pinkish dolomite and Mg-calcites, most prominent in the Alpha-Mendeleev Ridge region and on the Morris Jesup Rise8. Moreover, a lower boundary of the presumed B1 layer is indicated by such a dashed line8. White values = radiocarbon measurements of C. wuellerstorfi in ka 14C ages.

Fig. 2: Documentation of radiocarbon measured specimens and their assigned PI from 22.5 cm sediment depth in BC PS72/413-3.
figure 2

a PI_1 specimens pristine and translucent white, from this sample and other samples in the core section 20.5–23.5 cm revealed a GI age. b PI_2 specimens with an off-white shell lacking overgrowth reveal an MIS2 age. c PI_3 specimens with an off-white shell showing overgrowth reveal an MIS3 age.

In their comment, Löwemark and Singh raise the concern that the observed radiocarbon offset between Neogloboquadrina pachyderma shells of different diagenetic overprint in glacial sediments is likely an artifact and that the observed age offsets are rather due to Zoophycos bioturbation. In our cores, until becoming infinite (>39.2 ka) the age offset between PI_2 and PI_3 increases with depth, as does the thickness of overgrowth estimated by SEM images.

Löwemark and Singh presume that pristine foraminifera were distributed downcore in Zoophycos traces and measured as PI_2 specimens and that this could have been avoided by working on the slabs used for radiography12. As has been shown by, e.g., Küssner et al.13, this is a good method to elucidate the potential impact of large bioturbation structures on proxy measurements in foraminifera-rich sample13. However, in our case, this was not possible because the slabs and samples were processed 30 and 10 years before our current analyses for cores PS2185-6 and PS72/413-3, respectively, and the slabs and PS2185-6 sediments are no longer existent. Therefore, allochthonous foraminifera can not be avoided by sampling visually Zoophycos-free sediments. The thickness of x-ray sediment slabs is 1 cm, whereas, the mean thickness of a foraminifera is only 100–150 µm. As radiography integrates the sediment density of the 1-cm thick slab, one still has a high chance to include bioturbated foraminifera when sampling non-laminated and Zoophycos-free sediments. Central Arctic Ocean sediment cores are usually retrieved from ridges/seamounts where the sedimentation rate is very low and a lot of shallow to intermediate water depth foraminifera are deposited by drifting sea ice (see Fig. 29 of Wollenburg14) or icebergs15. Drifting icebergs further may erode sediments and re-deposit them nearby or elsewhere16. In addition to massive diagenetic shell changes, allochthonous and autochthonous shells are distributed relatively evenly, especially in the brown layers. It is, therefore, of great importance to keep an eye on both the lithology and the state of preservation when isolating foraminifera for radiocarbon dating. Our study shows that bioturbated PI_1 shells can be clearly distinguished from autochthonous PI_2 and PI_3 individuals due to their better preservation (Fig. 2). The similar preservation of PI_2 and PI_3 and the MIS2 age of PI_2 individuals further contrast with the brown Holocene to Greenland Interstadial 1 sediments in the Zoophycos traces. Finally, we would like to point out that at present, due to the strong diagenetic imprint on shells, the often intense bioturbation, and the many17 allochthonous components, a reliable age model in these cores must be supported by further independent stratigraphic methods (e.g., seawater derived Be isotopes17). We concur with Löwemark and Singh that bioturbation may bias Arctic records in specific instances, and we did our best to address and consider these effects in the original publication. However, our aim was to draw attention to the massive authigenic overgrowths on foraminiferal shells and their importance for proxy applications, especially radiocarbon dating, a problem that exists completely independently of bioturbation.