The cause of a rapid change in Atlantic Ocean circulation and northern cooling at the onset of Heinrich Stadial 1~18.5 ka is unclear. Previous studies have simulated the event using ice sheet and/or iceberg meltwater forcing, but these idealized freshwater fluxes have been unrealistically large. Here we use a different approach, driving a high-resolution drainage network model with a recent time-resolved global paleo-ice sheet reconstruction to generate a realistic meltwater forcing. We input this flux to the Hadley Centre Coupled Model version 3 (HadCM3) climate model without adjusting the timing or amplitude and find that an acceleration in northern ice sheet melting (up to~7.5 m/kyr global mean sea level rise equivalent) triggers a 20% reduction in the Atlantic Meridional Overturning Circulation. The simulated pattern of ocean circulation and climate change matches an array of paleoclimate and ocean circulation reconstructions for the onset of Heinrich Stadial 1, in terms of both rates and magnitude of change. This is achieved with a meltwater flux that matches constraints on sea level changes and ice sheet evolution around 19-18 ka. Since the rates of melting are similar to those projected for Greenland by 2200, constraining the melt rates and magnitude of climate change during Heinrich Stadial 1 would provide an important test of climate model sensitivity to future ice sheet melt. Plain Language Summary Atlantic Ocean circulation plays a key role in redistributing heataround Earth's surface, and thus has an important influence on our climate. Because of this, sudden shifts in Atlantic Ocean circulation can drive rapid climate changes. One such example is at the onset of "Heinrich Stadial 1", 18.5 thousand years ago, when geological records show that Atlantic circulation weakened and the Northern Hemisphere cooled while the Southern Hemisphere warmed. At the time, huge ice sheets (several kilometers thick) covered much of North America and northern Europe. Climate model results suggest that the freshwater produced by these melting ice sheets is responsible for weakening the Atlantic Ocean circulation and triggering the abrupt climate changes captured in the geological records. This result helps to elucidate the complex interaction between ice sheets, ocean circulation, and climate, and how these interactions can lead to sudden shifts in climates of the past and, potentially, the future. Indeed, the rate of melting we adopt in the present model is comparable to the melting projected for the Greenland Ice Sheet by 2200.

Low-resolution, complex general circulation models (GCMs) are valuable tools for studying the Earth system on multimillennial timescales. However, slowly evolving salinity drifts can cause large shifts in climatic and oceanic regimes over thousands of years. We test two different schemes for neutralising unforced salinity drifts in the FAMOUS GCM: surface flux correction and volumetric flux correction. Although both methods successfully maintain a steady global mean salinity, local drifts and subsequent feedbacks promote cooling (≈ 4 °C over 6000 years) and freshening (≈ 2 psu over 6000 years) in the North Atlantic Ocean, and gradual warming (≈ 0.2 °C per millennium) and salinification (≈ 0.15 psu per millennium) in the North Pacific Ocean. Changes in the surface density in these regions affect the meridional overturning circulation (MOC), such that, after several millennia, the Atlantic MOC (AMOC) is in a collapsed state, and there is a strong, deep Pacific MOC (PMOC). Furthermore, the AMOC exhibits a period of metastability, which is only identifiable with run lengths in excess of 1500 years. We also compare simulations with two different land surface schemes, demonstrating that small biases in the surface climate may cause regional salinity drifts and significant shifts in the MOC (weakening of the AMOC and the initiation then invigoration of PMOC), even when the global hydrological cycle has been forcibly closed. Although there is no specific precursor to the simulated AMOC collapse, the northwest North Pacific and northeast North Atlantic are important areas that should be closely monitored for trends arising from such biases.

<p><strong>Abstract.</strong> Constraining ocean circulation and its temporal variability is crucial for understanding changes in surface climate and the carbon cycle. Radiocarbon (<sup>14</sup>C) is often used as a geochemical tracer of ocean circulation, but interpreting ∆<sup>14</sup>C in geological archives is complex. Isotope-enabled models enable us to directly compare simulated ∆<sup>14</sup>C values to Δ<sup>14</sup>C measurements and investigate plausible mechanisms for the observed signals. We have added three new tracers (water age, abiotic <sup>14</sup>C, and biotic <sup>14</sup>C) to the ocean component of the FAMOUS General Circulation Model to study large-scale ocean circulation and the marine carbon cycle. Following a 10 000 year spin-up, we prescribed the Suess effect (the isotopic imprint of anthropogenic fossil fuel burning) and the bomb pulse (the isotopic imprint of thermonuclear weapons testing) in a transient simulation spanning 1765 to 2000 CE. To validate the new isotope scheme, we compare the model output to direct ∆<sup>14</sup>C observations in the surface ocean (pre-bomb and post-bomb) and at depth (post-bomb only). We also compare the timing, shape and amplitude of the simulated marine bomb spike to ∆<sup>14</sup>C in geological archives from shallow-to-intermediate water depths across the North Atlantic. The model captures the large-scale structure and range of ∆<sup>14</sup>C values (both spatially and temporally) suggesting that, on the whole, the uptake and transport of <sup>14</sup>C are well represented in FAMOUS. Differences between the simulated and observed values arise due to physical model biases (such as weak surface winds and over-deep North Atlantic Deep Water), demonstrating the potential of the <sup>14</sup>C tracer as a sensitive, independent tuning diagnostic. We also examine the importance of the biological pump for deep ocean <sup>14</sup>C concentrations and assess the extent to which <sup>14</sup>C can be interpreted as a ventilation tracer. Comparing the simulated biotic and abiotic δ<sup>14</sup>C, we infer that biology has a spatially heterogeneous influence on <sup>14</sup>C distributions in the surface ocean (between 18 and 30 ‰), but a near constant influence at depth (≈ 20 ‰). Nevertheless, the decoupling between the simulated water ages and the simulated <sup>14</sup>C ages in FAMOUS demonstrates that interpreting proxy ∆<sup>14</sup>C measurements in terms of ventilation alone could lead to erroneous conclusions about palaeocean circulation. Specifically, our results suggest that ∆<sup>14</sup>C is only a faithful proxy for water age in regions with strong convection; elsewhere, the temperature dependence of the solubility of CO<sub>2</sub> in seawater complicates the signal.</p>

Abstract. Ocean circulation and the marine carbon cycle can be indirectly inferred from stable and radiogenic carbon isotope ratios (δ13C and Δ14C, respectively), measured directly in the water column, or recorded in geological archives such as sedimentary microfossils and corals. However, interpreting these records is non-trivial because they reflect a complex interplay between physical and biogeochemical processes. By directly simulating multiple isotopic tracer fields within numerical models, we can improve our understanding of the processes that control large-scale isotope distributions and interpolate the spatiotemporal gaps in both modern and palaeo datasets. We have added the stable isotope 13C to the ocean component of the FAMOUS coupled atmosphere–ocean general circulation model, which is a valuable tool for simulating complex feedbacks between different Earth system processes on decadal to multi-millennial timescales. We tested three different biological fractionation parameterisations to account for the uncertainty associated with equilibrium fractionation during photosynthesis and used sensitivity experiments to quantify the effects of fractionation during air–sea gas exchange and primary productivity on the simulated δ13CDIC distributions. Following a 10 000-year pre-industrial spin-up, we simulated the Suess effect (the isotopic imprint of anthropogenic fossil fuel burning) to assess the performance of the model in replicating modern observations. Our implementation captures the large-scale structure and range of δ13CDIC observations in the surface ocean, but the simulated values are too high at all depths, which we infer is due to biases in the biological pump. In the first instance, the new 13C tracer will therefore be useful for recalibrating both the physical and biogeochemical components of FAMOUS.

No abstract