Freshwater routing in eddy-permitting simulations of the last deglacial: the impact of realistic freshwater discharge

Abstract. Freshwater, in the form of glacial runoff, is hypothesized to play a critical role centennial- millennial-scale climate variability, such as Younger Dryas and Dansgaard–Oeschger events, but this relationship not straightforward. Large-scale runoff Meltwater Pulse 1a (MWP1a), are always temporally proximal subsequent large-scale cooling. Moreover, typical design hosing experiments that support tends artificially amplify response. This study explores impact limitations representation conventional “hosing” simulations has on our understanding by examining where coastally released freshwater transported when it reaches ocean. We particularly focus (1) injection directly over sites deep-water formation (DWF) rather than at locations (i.e. hosing), (2) excessive volumes (often factor 5), (3) use present-day (rather palaeo) ocean gateways. track routing glaciologically constrained from four different inferred suite eddy-permitting using Massachusetts Institute Technology General Circulation Model (MITgcm) under both open closed Bering Strait conditions. Restricting forcing values realistic ranges results less spreading across North Atlantic indicates anomalies DWF depend strongly geographical location meltwater input. In particular, into Gulf Mexico generates very weak signal regions result entrainment turbulent Stream. contrast, Arctic with an or Eurasian ice sheet found generate largest salinity GIN (Greenland–Iceland–Norwegian) seas region respectively. Experiments show open, Mackenzie River source exhibits more twice much freshening closed. Our illustrate applying even “realistic” amounts still overestimates amount terrestrial reaching regions. Given simulated anomaly distributions lack reconstructed during Bølling–Allerød, majority American contribution MWP1a was routed through River.