1 Introduction
Freshwater originating from the Antarctic Ice Sheet (AIS) enters the ice shelf cavities and marginal seas, lowering local salinities and potentially hindering the formation of Antarctic Bottom Water (AABW) [Silvano et al. , 2018; Wijk and Rintoul , 2014]. AABW formation is one of the main drivers of the Global Meridional Overturning Circulation (GMOC) [Gordon 1986], a large-scale ocean circulation system that connects all ocean basins and controls global climate stability [Broecker , 1991]. Thus, by changing the rate of AABW formation, AIS melting can significantly affect the state of the global climate.
AABW is mainly formed through the mixing of Antarctic ice shelf waters and Circumpolar Deep Water (CDW), which flows towards the Antarctic coast and onto the continental shelf where it enters ice shelf cavities [e.g., Talley 2013 ]. Under the ice shelves, CDW mixes with Dense Shelf Water, creating dense waters that flow through the continental shelf, forming AABW through mixing along the shelf overflow path [Carmack and Foster, 1975; Foster and Carmack, 1976 ]. Ice melting also occurs under ice shelves, releasing freshwater and thus decreasing salinities along the Antarctic coast. Hence, increasing AIS melting can increase the buoyancy of shelf waters, hindering the formation of AABW [Wijk and Rintoul , 2014]. Alternatively, AABW can be formed offshore by deep convection in open-ocean polynyas [e.g., Killworth , 1983], a process commonly present in models [Aguiar et al , 2017; Azaneu et al , 2013], but only observed a few times since satellites started monitoring sea ice in the Southern Ocean in late 1972 [Campbell et al , 2019;Gordon 1978].
Two main processes release meltwater from AIS to the Southern Ocean. First, while circulating within ice-shelf cavities, warm CDW exchanges heat with the overlying ice and induces melting on the ice shelf base, a process known as basal melting [Jenkins et al., 2001 ]. Second, ice can detach from the ice shelves creating icebergs - a process referred to as calving [e.g., Joughin and MacAyeal, 2005 ]. Detached icebergs are advected by surface winds and ocean currents [e.g., Wagner et al., 2017 ], releasing freshwater along their advective path. Besides their potential impact on AABW formation, both ice shelf basal melting and calving are thought to increase sea-ice production and thus brine rejection, which can preclude the direct surface freshening by AIS melting [Bintanja et al ., 2013]. However, the most widely used Earth System Models do not explicitly simulate these ice melting processes, making it difficult to understand their impact on AABW formation and sea ice production [Jongma et al ., 2013].
Coupling ice sheet models to climate models to simulate AIS melting is technically complex and requires considerable computational effort [Nowicki and Seroussi, 2018 ]. An alternative approach is to add virtual freshwater fluxes in the Southern Ocean, a protocol suggested by the Ocean Model Intercomparison Project Phase 1 (OMIP1) [Farneti et al, 2015; Griffies et al, 2009 ]. Embedded model configurations often follow OMIP1 protocol by distributing AIS melting equally over every surface grid cell of the ocean model along the Antarctic coast, in a uniform flux configuration (Fig 1a). However, this uniform flux misrepresents the spatially varying calving and basal melting of ice shelves [Rignot et al., 2019 ], and therefore could impose salinity biases under ice shelf cavities, changing the salinity of AABW and its source waters [Jongma et al, 2009 ]. Since most state-of-the-art climate models misrepresent AABW properties [Heuzé et al, 2013 ], assessing the role of the spatial distribution of AIS melting in determining AABW properties is a useful step to improve the representation of the AABW in model simulations. Furthermore, a significant freshening of the AABW was reported over the last three decades [Purkey and Johnson, 2012; Anikulmar et al, 2021 ], with regional increases in AIS melting and calving possibly triggering this freshening [Fogwill et al, 2015 ]. This connection between enhanced regional AIS melting and AABW freshening makes it even more critical to assess how spatial variations in AIS melting alter AABW properties.
Therefore, this work assesses the impact of the freshwater distribution (uniform along the Antarctic coast versus spatially varying) on the AABW properties and transport in a global ocean and sea-ice coupled model, and further analyzes the potential response of AABW and the Antarctic sea ice to an enhanced AIS melting. Section 2 describes the methods, including the model, water mass definitions, and experimental design. Section 3 compares the surface and bottom responses to diverse spatio-temporal meltwater flux distributions. The main discussion and conclusions are provided in section 4.