At high flow rates the plume water is warmed to a lesser degree by the warm ambient water due to a larger volume of cold water entering the system. We will now analyse the combined effect of varying both S and Q, and also consider the depth at which the temperature maximum occurs. The plume’s mixing with warmer ambient waters (especially the Atlantic Water) warms the initially cold flow of dense water and also changes the depth distribution of temperature. For all model runs we determine the temperature maximum and depth of the temperature maximum found in the bottom model level at the end of each experiment. The results are plotted against S and Q to investigate the full range of forcing parameters for learn more all

model GSK1120212 in vitro runs. In Fig. 9 each experiment is marked by a black dot at a modelled combination of S and Q and the temperature

maximum (in Fig. 9(a)) and its depth (in Fig. 9(b)) are shaded as coloured contours that span the S-Q space. Fig. 9(a) shows that the magnitude of the temperature maximum (in °C) is primarily dependent on Q   and almost independent of S  , which confirms the interpretation of Fig. 8 for a wider range of forcing parameters. Cascades with low flow rates ( Q⩽0.02Sv) are warmed by the ambient water to 0.2 °C and above, while at higher flow rates ( Q⩾0.03Sv) the cold cascade lowers the temperature maximum below 0 °C. The flow rate dependence of the maximum bottom temperature in Fig. 9(a) can be explained by the different thermal capacity of the volume of plume water as Q changes, compared to the unchanged thermal capacity of the Atlantic Water. The salinity dependence of the depth of the temperature maximum in Fig. 9(b) is related to the salinity being the main driver of density at low temperatures. Plumes of lower salinity are thus less dense, causing them to advance downslope at slower speeds. A slowly descending plume remains in the Atlantic Layer for longer and

more AW is mixed into the plume. Hence more warm Atlantic water gets advected check details downslope, causing the temperature maximum to occur at deeper depths in experiments with low S. The mixing between the cold cascade and the warm ambient waters does not only lower the bottom-level temperature maximum, it also alters its depth which initially occurs within between 200 and 500 m at the start of each experiment. Fig. 9(b) shows that the depth of the temperature maximum has been displaced upslope (shallower than 400 m, shaded yellow) or downslope (deeper than 600 m, shaded blue) by the end of each experiment. In experiments where S⩽35.20S⩽35.20 the temperature maximum occurs at depths of 600 to 800 m while it remains at shallower depths of 200 to 400 m in experiments with S > 35.20. We conclude that the final depth of the temperature maximum is thus primarily dependent on the inflow salinity S. By prescribing a varying salinity at the overflow we are able to recreate (in Fig.