Glacial periods are characterized by lower atmospheric CO₂ concentrations than present, leading to global cooling and significant carbon storage in the deep ocean. While the Southern Ocean's role as a major CO₂ sink during glacial periods is established, the Arctic Ocean's contribution remains unclear. The prevailing paradigm assumes a mostly oxygenated Arctic Ocean during the late Quaternary. However, recent proposals suggest a freshwater-filled Arctic basin during certain marine isotope stages (MIS), prompting a reevaluation of Arctic paleohydrology. This study uses authigenic carbonates in glacimarine sediments from the Mendeleev Ridge to reconstruct the salinity and oxygenation of the Polar Deep Water (PDW) in the western Arctic Ocean during the last glacial period, aiming to clarify the Arctic Ocean's role in glacial CO₂ drawdown.

Previous research on Arctic Ocean paleohydrology has yielded conflicting geochemical evidence, partly due to discontinuous foraminifera records, which are poorly preserved during periods of freshening and anoxia. Studies focusing on the absence of manganese in glacial sediments suggested dysoxic bottom waters. The recent hypothesis of a freshwater-filled Arctic Ocean during MIS 6 and 4 (ref. ⁵) challenges the existing paradigm of an oxygenated Arctic Ocean and highlights the need for alternative proxies to reconcile inconsistencies in paleoceanographic reconstructions. This study addresses this need by utilizing authigenic carbonate records, offering a potentially continuous record of PDW characteristics.

This study analyzed authigenic carbonates from a 39-cm-long sediment core (PS72/410-1) collected from the Mendeleev Ridge in the western Arctic Ocean. Authigenic carbonates, formed in situ, provide direct information on the prevailing oceanographic conditions of the overlying PDW. The researchers examined carbonate crystal textures using scanning electron microscopy (SEM) and determined mineralogy (calcite, Mg-calcite, aragonite) using crystal shapes and energy-dispersive X-ray spectroscopy (EDAX). Stable isotope analysis (δ¹⁸O and δ¹³C) was conducted using a gas mass spectrometer to assess the isotopic composition of the carbonates, and magnesium concentrations were analyzed using inductively coupled plasma atomic emission spectroscopy (ICP-AES). The age of the sediment core was established using a revised lithostratigraphy based on radiocarbon ages and color reflectance values, correlating with neighboring cores. The researchers used various statistical analyses to determine correlations between data sets.

Authigenic calcites were found throughout the core, except for a few intervals. Two intervals showed relatively higher carbonate contents. Calcite crystals exhibited various growth textures, indicating inorganic origin, including bundles of platy crystals and individual crystals. Aragonite crystals mostly showed fibrous textures, sometimes as overgrowths on calcite. The presence of authigenic calcites in the core-top sediment suggests persistent calcite saturation in the ambient seawater over the last 76 kyr. The geochemical analysis of authigenic calcites showed that during stadial periods (MIS 4-3 and MIS 2), low-magnesium calcites (LMC) predominated, contrasting with high-magnesium calcites during the Holocene and interstadial within MIS 3. The occurrence of LMC during stadials reflects a significant reduction in Mg/Ca ratios in the PDW, attributable to massive freshwater sinking, rather than mixing with surrounding oceans or hydrocarbon seepage. This is further supported by lower strontium (Sr) contents in LMC during stadials. Calcite δ¹⁸O values show a wide range (~11‰), primarily reflecting freshwater intrusion rather than temperature variability. The δ¹⁸O-depleted freshwater likely originated from the Eurasian and East Siberian Ice Sheets (MIS 4-3) and the Laurentide Ice Sheet (MIS 2). The PDW freshening is consistent with previous studies based on ²³⁰Thₑₓ but differs slightly in timing and extent of freshening (brackish rather than fully fresh). Positive δ¹³C signatures during MIS 2 are higher than those of dissolved inorganic carbon (DIC) and benthic foraminifera, suggesting anoxic conditions and methanogenesis. The anoxic conditions are attributed to oxygen-poor meltwater intrusion, limited air-sea exchange under thick sea ice, and inhibited Atlantic water inflow. The presence of calcareous foraminifera under anoxic conditions is explained by episodic ecosystem recovery due to intermittent oxygen and salt supply through hyperpycnal flows and/or limited Atlantic inflow. Slight positive δ¹³C excursions during other intervals also suggest persistent anoxic environments. In contrast, warmer periods showed abundant benthic foraminifera, reflecting deep convection and oxygen supply via brine formation. The inconsistency with authigenic uranium enrichment is likely due to reduced uranium concentrations in seawater from freshening.

The findings demonstrate that the PDW in the western Arctic Ocean was predominantly brackish and anoxic during stadial periods (MIS 4-3 and MIS 2), suggesting a significant role of the western Arctic Ocean as a carbon sink during glacial periods. The hyperpycnal subglacial meltwater acted as a primary carrier of glacially derived carbon to the deep ocean, preventing its release into the atmosphere. This contrasts with previous assumptions of an oxygenated Arctic Ocean and highlights the need to consider the Arctic Ocean as a dynamic player in the global carbon cycle. The study's findings are consistent with observations of oxygen depletion in other glacial deep ocean regions, further emphasizing the ubiquitous nature of oxygen depletion events in the glacial deep ocean.

This study provides strong evidence for brackish and anoxic conditions in the western Arctic Ocean's PDW during stadial periods of the last glacial cycle. The influx of oxygen-poor meltwater from extensive ice sheets significantly influenced the PDW's chemistry and created a substantial carbon reservoir. Future research should focus on quantitative O₂ reconstructions across the entire Arctic Ocean to better understand greenhouse gas behavior across glacial-interglacial cycles and fully quantify the Arctic's role in the global carbon cycle.

The study primarily focuses on the western Arctic Ocean, limiting the generalizability of findings to the entire Arctic Ocean. The age model relies on a revised lithostratigraphy and correlations with neighboring cores, introducing uncertainties. Furthermore, while the study provides compelling evidence for anoxia and methanogenesis, the exact mechanisms driving these processes and their relative contribution require further investigation. The impact of sediment-laden hyperpycnal flows needs further assessment in terms of depth penetration and spatial extent.