For better projections of sea level rise, two things are needed: an improved understanding of the contributing processes and their accurate representation in climate models. A major process is basal melting of ice shelves and glacier tongues by the ocean, which reduces ice sheet stability and increases ice discharge into the ocean. We study marine melting of Greenland’s largest floating ice tongue, the 79° North Glacier, using a high-resolution, 2D-vertical ocean model. While our fjord model is idealized, the results agree with observations of melt rate and overturning strength. Our setup is the first application of adaptive vertical coordinates to an ice cavity. Their stratification-zooming allows a vertical resolution finer than 1 m in the entrainment layer of the meltwater plume, which is important for the plume development. We find that the plume development is dominated by entrainment only initially. In the stratified upper part of the cavity, the subglacial plume shows continuous detrainment. It reaches neutral buoyancy near 100 m depth, detaches from the ice, and transports meltwater out of the fjord. Melting almost stops there. In a sensitivity study, we show that the detachment depth depends primarily on stratification. Our results contribute to the understanding of ice–ocean interactions in glacier cavities. Furthermore, we suggest that our modeling approach with stratification-zooming coordinates will improve the representation of these interactions in global ocean models. Finally, our idealized model topography and forcing are close to a real fjord and completely defined analytically, making the setup an interesting reference case for future model developments.

A realistic numerical model is used to study the circulation and mixing of the Salish Sea, a large, complex estuarine system on the United States and Canadian west coast. The Salish Sea is biologically productive and supports many important fisheries but is threatened by recurrent hypoxia and ocean acidification, so a clear understanding of its circulation patterns and residence times is of value. The estuarine exchange flow is quantified at 39 sections over three years (2017-2019) using the Total Exchange Flow method. Vertical mixing in the 37 segments between sections is quantified as opposing vertical transports: the efflux and reflux. Efflux refers the rate at which deep, landward-flowing water is mixed up to become part of the shallow, seaward-flowing layer. Similarly, reflux refers to the rate at which upper layer water is mixed down to form part of the landward inflow. These horizontal and vertical transports are used to create a box model to explore residence times in a number of different sub-volumes, seasons, and years. Residence times from the box model are generally found to be longer than those based on simpler calculations of flushing time. The longer residence times are partly due to reflux, and partly due to incomplete tracer homogenization in sub-volumes. The methods presented here are broadly applicable to other estuaries.