Arctic drainage of Laurentide Ice Sheet meltwater throughout the past 14,700 years

The study addresses when and how meltwater from the retreating Laurentide Ice Sheet (LIS) drained along its major outlets during the last deglaciation, and whether Arctic routing via the Mackenzie River contributed to abrupt climate events such as the Younger Dryas. After the Last Glacial Maximum, meltwater routing occurred south via the Mississippi, east via the St. Lawrence, north via Hudson Bay/Strait, and northwest via the Mackenzie. While early deglacial discharge to the Gulf of Mexico is well established, routing during and after the onset of the Bølling–Allerød warm interval and around the Younger Dryas remains debated, especially the feasibility of St. Lawrence routing versus northwestern overflow through the Fort McMurray–Athabasca spillway. Terrestrial evidence is sparse or ambiguous due to erosion by floods, poor radiocarbon datability, and incomplete field records. Marine archives, though less sensitive to small freshwater changes, can preserve continuous, well-dated signals of extreme events. The authors aim to reconstruct freshwater routing into the Arctic Ocean since the Bølling–Allerød using authigenic Pb isotopes in Beaufort Sea and Amundsen Gulf sediments, to test hypotheses about Lake Agassiz drainage via the Mackenzie at the Younger Dryas onset and to build an integrated drainage chronology.

Prior work identified main LIS drainage routes (Mississippi, St. Lawrence, Hudson Bay/Strait, Mackenzie) and proposed that a Lake Agassiz outburst via the St. Lawrence triggered the Younger Dryas (Broecker 1989). However, geomorphic constraints suggest eastern routing may have been blocked during the Younger Dryas (e.g., elevated surfaces, moraines, LIS blockage), and terrestrial indicators of northwestern routing are incomplete or debated. Kennett and Shackleton (1975) first used low planktonic δ18O in the Gulf of Mexico to indicate deglacial meltwater, and similar δ18O lows were found off the Mackenzie Delta. Yet δ18O has limitations (foraminiferal absence in fresh surface waters, mixing of individuals, lack of source specificity). Authigenic Pb isotopes captured in Fe-Mn oxyhydroxide coatings have been shown to sensitively trace weathering/runoff signals and provenance, with combined detrital/authigenic phases discriminating sediment source versus runoff chemistry (e.g., Kurzweil et al. 2010; Gutjahr et al. 2009). Numerous studies reconstruct LIS evolution and Lake Agassiz history, including large lake-level drops around the Younger Dryas, possible Athabasca spillway activation, Amundsen Gulf ice-stream collapse, and timing of eastern outlet opening, but a coherent drainage chronology across all outlets has remained elusive.

Study area and cores: Three sites cored during USCGC Healy cruise HLY1302 along the Canadian Arctic margin were analyzed: JPC-15/27 (71.10°N, 135.13°W, 687 m) in the central Beaufort margin, JPC-19 (71.29°N, 126.28°W, 442 m) in the Amundsen Gulf, and JPC-9 (70.58°N, 142.42°W, 394 m) west of the Mackenzie Delta. JPC-15 and JPC-27 were spliced using magnetic susceptibility to form a composite (JPC-15/27). The records span from late Heinrich Stadial 1 to the Holocene. Age models: Radiocarbon dating on planktonic (and in places mixed planktonic/benthic) foraminifera was performed at the LARA lab (University of Bern) and calibrated with CALIB 8.2 using Marine20 and AR values consistent with Keigwin et al. (2018). JPC-15/27 has 14 14C dates back to ≥15 ka; JPC-9 includes five 14C dates; JPC-19 includes four new dates. Geochemical extractions and isotope analysis: The authigenic Fe-Mn oxyhydroxide phase was leached from ~0.5 g sediment with a weak reductive solution (0.005 M hydroxylamine hydrochloride, 1.5% acetic acid, 0.003 M Na-EDTA, pH 4) for ~30 s. Residual authigenic coatings were removed from the detrital fraction with a tenfold stronger solution (0.05 M hydroxylamine hydrochloride, 15% acetic acid, 0.03 M Na-EDTA, pH 4) for ≥12 h. Detrital sediments were homogenized and ~50 mg digested (microwave) in concentrated HNO3 and HBF4. Al/Pb ratios were monitored to assess detrital contamination in authigenic extracts. Pb was isolated by AG1-X8 ion-exchange chromatography. Pb isotope ratios (206Pb/204Pb, 208Pb/204Pb) were measured by MC-ICP-MS (Thermo Neptune Plus) at GEOMAR with Tl-doping mass bias correction (target Pb/Tl ≈ 4). Accuracy (2SD) on USGS NOD-A-1 was ~0.0055 for 206Pb/204Pb and ~0.015 for 208Pb/204Pb; blanks ~500 pg (<0.1% of sample). Elemental concentrations (Li, Al, P, Ca, Ti, Mn, Fe, Sr, Ce, Nd, Pb, Th, U) were measured by ICP-MS (Agilent 7500), reproducibility better than 2% (2SD). Rock magnetic analyses: U-channel samples from JPC-15/27 were measured on a 2G SRM. Anhysteretic remanent magnetization (ARM) was imparted with 100 mT peak AF and 0.05 mT bias, measured before (ARM0mT) and after 30 mT AF (ARM30mT). Magnetic susceptibility and ARM track ferrimagnetic mineral concentration; ARM30mT/ARM0mT indicates magnetic coercivity and relative fine-grained magnetite content. Together with grain-size and geochemical proxies (e.g., Ca/Sr), these data help distinguish provenance versus particle-size effects. Complementary data: Planktonic foraminiferal δ18O (Neogloboquadrina pachyderma s.) from JPC-15/27 were compared with Pb isotopes, and regional records (Gulf of Mexico, Laurentian Fan, Hudson Strait) were compiled to build a drainage chronology.

• Continuous Arctic routing: Authigenic Pb isotopes at JPC-15/27 indicate persistent meltwater supply to the Arctic via the Mackenzie valley since the onset of the Bølling–Allerød (~14.7 ka), marked by an initial shift to more radiogenic 206Pb/204Pb and 208Pb/204Pb at ~1330 cm depth. • Strongest Lake Agassiz outflow: The most radiogenic authigenic Pb excursion occurs just before the Younger Dryas at ~13.1 ka (530 cm), predating the planktonic δ18O excursion and providing direct evidence for a major Lake Agassiz overflow via the Athabasca spillway–Mackenzie route near the BA–YD transition. • Multiple subsequent pulses: Additional radiogenic authigenic Pb excursions occur during the Younger Dryas and early Holocene at ~12.0 ka, ~11.2 ka, and a smaller event starting ~9.7 ka, suggesting repeated overflows; amplitudes diminish through time. • Decoupled sediment vs runoff signals: During these events, detrital Pb isotopes do not mirror authigenic excursions, indicating that freshwater (runoff chemistry) and terrigenous sediment sourcing were often decoupled, likely due to sediment trapping in intermediate proglacial lakes (e.g., Mackenzie) and transport constraints. • Mid-BA Amundsen Gulf event: JPC-19 registers a pronounced, ≥1 kyr radiogenic excursion in both detrital and authigenic Pb (the “Amundsen Gulf event”), reflecting major freshwater/iceberg outflow tied to the Amundsen Gulf Ice Stream collapse; subtle influence of this event is potentially seen at JPC-15/27. • Holocene stabilization: After ~8.4–8.5 ka, authigenic Pb ratios at JPC-15/27 stabilize near 206Pb/204Pb ≈ 19.3 and 208Pb/204Pb ≈ 39.1, indicating establishment of modern-like Mackenzie drainage with diminished glacial-lake influence. • Lateral gradients: JPC-9 shows similar but weaker authigenic radiogenic excursions during YD–early Holocene, consistent with preferential eastward transport by the shelfbreak current toward central Beaufort Sea. • Integrated drainage chronology: Five deglacial phases are delineated. (1) Early BA (14.7–13.8 ka): increased GOM discharge (south), first Arctic signal from northern glacial lakes and Amundsen/Mackenzie sectors. (2) Late BA (13.8–13.1 ka): dramatic increase of southern outflow, first resolvable signal at Laurentian Fan. (3) BA–early YD (13.1–12.4 ka): opening of Athabasca spillway; dominant Arctic routing; southern discharge wanes. (4) YD–early Holocene (12.4–8.6 ka): concurrent eastern (St. Lawrence) and northern (Mackenzie) routing; Lake Agassiz overflows to Arctic at ~12.0 and ~11.2 ka. (5) 8.6–8.2 ka: final, most voluminous drainage of Lake Agassiz–Ojibway into Hudson Bay (Hudson Bay Ice Saddle collapse), terminating other outlets.

The Pb isotope records resolve the longstanding debate on LIS routing at the Younger Dryas onset by providing source-sensitive evidence that Lake Agassiz overflowed northward via the Athabasca spillway into the Mackenzie valley just before the YD, supporting Arctic freshening as a plausible trigger for the event. The earlier occurrence of the authigenic Pb peak relative to δ18O in the same core highlights the higher sensitivity of Pb isotopes to abrupt runoff changes. Subsequent smaller radiogenic excursions during the YD and early Holocene point to repeated, albeit diminishing, overflows or enhanced runoff events, consistent with independent gravel-bed evidence from the Mackenzie Delta. The divergence between authigenic and detrital Pb isotopic behavior during overflow phases indicates that terrigenous sediment delivery from the Lake Agassiz basin to the Arctic was muted relative to freshwater, likely due to sediment trapping in intermediate glacial lakes and the evolving catchment geometry; meanwhile, magnetic and grain-size data reflect transient increases in transport energy rather than wholesale sediment-provenance changes. Regionally, the Amundsen Gulf event demonstrates that deglacial dynamics in adjacent sectors could influence Beaufort margin signals, and differences between JPC-15/27 and JPC-9 suggest eastward-biased freshwater transport along the shelfbreak. Synthesizing with Gulf of Mexico, Laurentian Fan, and Hudson Strait records yields a coherent drainage chronology showing a shift from dominant southern routing early in deglaciation to strong northern routing at the BA–YD transition, then concurrent northern and eastern routing through the YD–early Holocene, and finally terminal eastern/northern discharge into Hudson Bay at 8.5–8.2 ka leading to modern drainage configurations. These findings refine our understanding of freshwater forcing pathways that can modulate ocean circulation and abrupt climate.

Authigenic and detrital Pb isotope records from Beaufort Sea and Amundsen Gulf sediments reveal continuous LIS meltwater routing to the Arctic via the Mackenzie since the Bølling–Allerød, with the strongest Lake Agassiz outflow just before the Younger Dryas. The data provide direct, source-sensitive evidence for a northwestern drainage route at the BA–YD transition, supporting Arctic freshening as a potential trigger of the Younger Dryas. Multiple subsequent, smaller overflows are detected during the YD and early Holocene, and post-8.5 ka stabilization of authigenic Pb isotopes marks the establishment of modern Mackenzie drainage following the Hudson Bay Ice Saddle collapse and Lake Agassiz–Ojibway drainage. The integrated chronology reconciles southern, eastern, and northern routing phases across Termination 1. Future work should better constrain the timing and magnitude of eastern (St. Lawrence) routing during the early Younger Dryas, quantify transport and mixing effects on Pb isotope signatures along pathways, and analyze Lake Agassiz sedimentary sequences to track evolving lake-water isotope compositions.

• Terrestrial geomorphic evidence for the Athabasca spillway and related routing is sparse or debated, limiting independent onshore corroboration. • Pb isotope formation and export reflect complex interactions among bedrock provenance, weathering intensity, exposure time, and hydrology; disentangling changing source contributions versus runoff magnitude remains uncertain. • The amplitude decline of later events cannot be uniquely attributed to reduced outflow volume versus changing source mixing. • Detrital and authigenic phases can be decoupled by transport and sediment trapping (e.g., in glacial lakes), complicating direct source-to-sink inferences. • The cause of the detrital Pb shift around ~8.5 ka remains unresolved. • Spatial coverage is limited to three marine cores; lateral gradients and pathway transformations (e.g., along-shelf currents) introduce uncertainty.