An analysis of mixing in the frontal zone of South and North Atlantic Central Water off North-West Africa

Multi-parameter mixing analysis is applied to 26 stations of Auftrieb 75, a joint expedition of R.R.S. Discovery and FS Meteor. North Atlantic Central Water (NACW) and South Atlantic Central Water (SACW) are represented by two water types each, and phosphate, silicate and nitrate are all used as an additional parameter with temperature and salinity, yielding three sets of independent solutions. Solutions from different parameters show similar vertical distributions but can differ considerably in absolute terms. This is believed to be caused by insufficient knowledge of nutrient values of pure NACW and SACW and by low range/precision ratios of nutrient determination. It is shown that isopycnal mixing analysis correctly describes the water mass distribution to lowest order, i.e. there is agreement between results from isopycnal and multi-parameter analysis at most stations within the error bounds which are inherent in the method and which are presently very high. Significant non-isopycnal mixing could be confirmed for only one station, situated on the continental slope in the area of the undercurrent of the Canary Current upwelling regime, and may be caused by active upwelling, increased lateral and vertical shear in the under-current, or internal waves interacting with the bottom. Five stations do not yield oceanographically acceptable solutions, and an attempt is made to include a coastal water mass in the analysis in order to improve the result for these stations, i.e. three-parameter analysis for NACW and SACW is replaced by four-parameter analysis for NACW, SACW and coastal water by combining temperature and salinity with phosphate and silicate or with nitrate and silicate simultaneously (the combination phosphate and nitrate is degenerate because both nutrients are linearly correlated). This indeed improves the results, indicating the presence of coastal water at those stations, but accumulation of errors inhibits quantitative conclusions.