JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 93, NO. C4, PAGES 3563-3569, APRIL 15, 1988 The Origin of Norwegian Sea Deep Water JAMES H. SWIFT OceanResearch Division,Scripps Institution of Oceanography, La Jolla, California KLAUS PETER KOLTERMANN ! Deutsches Hydrographisches lnstitut, Hamburg, Federal Republicof Germany A nearly homogeneouswater mass, the Norwegian Sea Deep Water, is found below 2000-m depth in the Norwegian and Lofoten basins of the Norwegian Sea. Recent observationsindicate that this water is a mixture of relatively cold and fresh Greenland Sea Deep Water with warmer, saltier Eurasian Basin Deep Water from the Arctic Ocean. We have found this mixture along the western and southern periphery of the Greenland Sea, near the level where the pressure-compensated densities of the parent water masses are equal. The along-isopycnal mixing produces a remarkably uniform water mass, which can be traced only a short distance away from its entry into the Norwegian Sea through gaps in the mid-ocean ridge north of Jan Mayen Island. Direct measurements of flow through these gaps confirm motion in the proper sense to accomplish this connection. INTRODUCTION When Helland-Hansen and Nansen [1909] discussed the "cold, heavy bottom-water" of the Norwegian Sea, they were concernedwith the origin of one water mass,in one geograph- ical domain: their Norwegian Sea included the waters between Greenland and Norway, or today's Greenland, Iceland, and Norwegian seas(see Figure 1 for present nomenclature), and in their view a single mass of cold water filled the abyss. On the basis of observations of nearly identical surface and bottom water characteristics in the vicinity of 75øN and 0 ø longitude (central Greenland Sea)and of exceptionally cold bottom waters there, they stated that [Helland-Hansen and Nansen, 1909, p. 324] "the formation of cold, heavy bottom- water was here directly observedon the very sea-surface" and that "the question of the process of its formation is thus finally settled beyond all doubt." As more has been learned of the nature of the Greenland, Iceland, and Norwegian seas, it has become clear that Helland-Hansen and Nansen had seen only part of the problem, and so held only a portion of the solu- tion, namely, there is at least one other distinct deep water mass in the domain. Although Helland-Hansen and Nansen's concept of the origin of Greenland Sea Deep Water (in effect, the formation event they described)remains a principal com- ponent of recent thinking on the process, that actual forma- tion event has yet to be unambiguously observed by today's oceanographers. Thus despite Helland-Hansen and Nansen's certainty, the study of the origin and circulation of the deep waters in this region has retained several fundamental problems, and one issue in particular has remained most puzzling: What is the origin of the deep water in the Norwegian and Lofoten basins of the Norwegian Sea?This paper presents the evidence which led us to the answer to this question. • Now at World Ocean Circulation Experiment International Plan- ning Office, Deacon Laboratory, Institute of Oceanographic Sciences, Godaiming, Surrey, England. Copyright 1988 by the American GeophysicalUnion. Paper number 8C0017. 0148-0227/88/008C-0017505.00 DEEP WATER MASSES Swift and Aa•laard [1981] used salinity observations to slightly expand Metcalf's [1960] and Aa•laard and Coach- man's[1968] temperature-based partition of the Greenland, Iceland, and Norwegian sea deep waters into two types: colder, fresherGreenland Sea Deep Water or GSDW (typi- cally 0 - - 1.30øC, S = 34.89PSU (practical salinity units, per mille salt by weight)) and warmer,saltierNorwegian SeaDeep Water, or NSDW (typically 0 =- 1.05øC, S = 34.91 PSU). We show in Table 1 the deep water characteristics averaged over the principal basinsholding the deep waters. In every respect, GSDW emerges as a relatively young water mass, certainlyin more intimate contactwith air-seaexchange pro- cesses than NSDW. In fact, the GSDW ventilation-related characteristics representan open ocean global extreme for deep waters;hence GSDW is one of the strongest potential bottom water candidates for formation tied to open ocean deep convection processes. While we note that the O-S charac- teristics of GSDW have yet to be unambiguously observed at the winter sea surface above the major basins, we do have reason to suspect that this is not so muchthe resultof GSDW characteristics' never being found at the sea surface (in which case GSDW would be formed by subsurface processes) as it is the result of our having accumulatedinsufficient midwinter data (see also Clarke et al. l-1986]). The characteristics of NSDW are markedly different from those of GSDW, and there are no observationsof its forma- tion at the seasurface independently of GSDW. For example, it cannot simply be an older GSDW representing different climatic conditions at formation time, because its burden of anthropogenic tracers increased markedlyfrom 1972 to 1981 without any marked change in its hydrographic characteristics (compare, for example, $rnethie et al. [1986]). Because anthro- pogenic tracers werefoundfirst in GSDW, then in NSDW, it is likely that NSDW is at leastin part modifiedGSDW. But until recently it wasnot understood how the transformation of GSDW to NSDW could occur. For instance, at no level in the central Greenland Sea is the water as salty as NSDW. There- fore a source of dense, salty water external to the Greenland Seais required. In this paperwe choose to simply accept, for want of hard evidence of the mechanism of its formation, the 3563