The hypothesis that meltwater pulses from the retreating Laurentide Ice Sheet (LIS) weakened the Atlantic Meridional Overturning Circulation (AMOC) and caused abrupt climate shifts during the last deglaciation has been widely discussed. While geomorphic evidence supports numerous meltwater discharge events coinciding with cool periods, uncertainties about flood characteristics (volume, duration) hinder the direct attribution of specific events to specific climate shifts. Recent modeling emphasizes the importance of meltwater input location and duration in influencing AMOC response. This study uses numerical modeling to test the hypothesis that Hudson Valley meltwater floods triggered the Intra-Allerød Cold Period (IACP), a cold period preceding the Younger Dryas. Understanding the impact of meltwater inputs on ocean circulation is crucial, especially with accelerating ice loss from Greenland and Antarctica increasing freshwater input into the oceans. The IACP, a cold period occurring towards the end of the Bølling-Allerød interstadial, is marked by a negative isotope excursion in Greenland ice core records between 13,311 and 13,099 yr b2k (before 2000 CE). The drainage of Glacial Lake Iroquois around 13,300 cal BP, resulting in a ~700 km³ flood down the Hudson Valley, and a subsequent larger flood (~2500 km³), has been proposed as a potential trigger for the IACP's AMOC weakening. The study aims to evaluate the impact of these flood events on the AMOC.

Previous research hypothesized a link between meltwater pulses from the Laurentide Ice Sheet and AMOC weakening, leading to abrupt climate changes during the last deglaciation. Studies have shown a temporal correlation between meltwater discharge events and centennial-to-millennial-scale cooling periods, but the exact mechanisms remain debated. The role of meltwater flood characteristics, such as volume and duration, and the importance of the input location, have been highlighted in recent modeling studies. The Intra-Allerød Cold Period (IACP), a significant cold event, has been linked to the drainage of Glacial Lake Iroquois, resulting in large meltwater floods down the Hudson Valley. However, the exact causal relationship and the sensitivity of AMOC to these events remain uncertain, underscoring the need for further investigation using high-resolution modeling.

This study employs a high-resolution (1/6°; 18-km) eddy-resolving global configuration of the coupled Massachusetts Institute of Technology general circulation model (MITgcm). The model simulates conditions 13,300 years ago with sea-level 75 m lower and uses modern atmospheric boundary conditions to represent the relatively warm Allerød period. A suite of experiments was designed to assess the sensitivity of AMOC to various meltwater discharge scenarios from the Hudson Valley, considering different combinations of flood duration (1 month and 1 year), frequency (single and multiple events), and volume, encompassing the estimated uncertainties. The "Realistic" scenario involved two floods of 700 km³ and 2500 km³, spaced 10 years apart. Other experiments manipulated the total volume (3700 km³) by changing flood duration and frequency. The study also considered the presence or absence of a continuous background meltwater flux of 0.05 Sv from the Hudson Valley. The model output includes sea surface salinity (SSS) anomalies, mixed layer depth (as a proxy for deep water formation), and AMOC strength. The Labrador and Greenland Seas were analyzed as key regions for NADW formation.

The simulations reveal that in all experiments, regardless of flood magnitude, duration, recurrence interval, or background flux, the meltwater from the Hudson Valley was rapidly mixed into the interior ocean by the Gulf Stream. By the time the freshwater reached the subpolar North Atlantic, where AMOC processes occur, it was no longer detectable. There was no significant reduction in surface salinity in these regions, compared to the control simulation. Even the most extreme scenario (three 1-month duration floods) resulted in rapid dissipation of freshwater, preventing the formation of a lower-density surface layer that could inhibit NADW formation. Mixed layer depth, a proxy for deep water formation, showed variations of less than 4% from the control in all experiments. Mean AMOC strength deviated by less than 1% from the control and exhibited modern-day variability. The rapid dissipation of meltwater appears to be driven by the interaction with the Gulf Stream, the long transport distance, and eddy-induced mixing in the high-resolution model. The location of meltwater input is a critical factor in determining its impact on AMOC. In addition, even the "Realistic" scenario only showed a faint SSS anomaly migrating north, insufficient to weaken NADW formation. Experiments with varying flood durations (1 month and 1 year) and frequencies (single and multiple events, even with a total volume exceeding that typically associated with the Younger Dryas event) produced no significant AMOC reduction. Even the inclusion of a continuous background meltwater flux did not alter the findings. The results suggest that the meltwater floods from the Hudson Valley were likely insufficient to cause a significant weakening of the AMOC.

The findings suggest that several factors prevented the Hudson Valley meltwater floods from significantly influencing the AMOC. The rapid entrainment and diffusion of meltwater by the Gulf Stream resulted in its mixing into the interior ocean before reaching regions crucial for AMOC processes. The long transport distance, combined with the energetic Gulf Stream and the relatively small flood volumes, limited the freshwater's impact. The study reinforces the importance of meltwater input location and the downstream flow paths to areas of deep water convection in determining AMOC response. The high-resolution model realistically captures the rapid mixing and dissipation of meltwater, contrasting with some lower-resolution models showing more persistent AMOC reductions. The study also suggests that the total meltwater volume from the Hudson Valley floods, even in the most extreme scenario, was insufficient to cause the prolonged cooling observed during the IACP compared to larger scale events such as the Younger Dryas. The inclusion of background runoff, even at half the typically proposed threshold for AMOC weakening, did not significantly alter the results. The model's prescribed atmosphere excluded potential atmospheric circulation changes influenced by the LIS, suggesting future work needs to explore the combined influences.

The study's results demonstrate that meltwater from the Hudson Valley floods, even in extreme scenarios, was unlikely to be the primary cause of the IACP cooling. The rapid dissipation of the meltwater by the Gulf Stream, long transport distances, and the relatively small volume compared to events like the Younger Dryas, limited its impact on the AMOC. Future research should focus on investigating the combined effects of meltwater from multiple sources, exploring the interactions between atmospheric circulation changes and meltwater inputs, and considering other mechanisms (volcanism, sea-ice feedbacks) that might have contributed to abrupt climate shifts during deglaciation. The findings highlight the critical importance of considering the geographic location and hydrodynamics of meltwater inputs in assessing their potential to alter large-scale ocean circulation and climate.

The study uses modern atmospheric boundary conditions as a proxy for the Allerød period, which could influence the results. The model excludes some potentially important interactions, such as those between atmospheric circulation and meltwater inputs. The focus on the Hudson Valley floods does not encompass the contributions from other sources of meltwater input. The sensitivity of the findings to model resolution and parameterizations needs to be considered.