No abstract

Hydrographic data from the first phase of the Coordinated Eastern Arctic Experiment (CEAREX) are analyzed. The data consist of temperature and salinity measurements made by a ship-based conductivity-temperature-depth (CTD) instrument and by a drifting SALARGOS buoy. These data were collected in the autumn and early winter of 1988-1989 in the northern Barents Sea, mostly in ice-covered conditions and then across the marginal ice zone (MIZ). The data show that relatively warm, salty water of Atlantic origin is modified by air cooling and ice melting to produce lighter water that has properties identical to (lower) halocline water found in the Arctic Ocean. This occurs mostly at the MIZ and to a lesser degree within the ice pack itself. At the MIZ the halocline water subducts underneath the lighter meltwater that resides directly under the ice pack; geostrophic velocity calculations indicate that it then turns eastward and flows toward the Kara Sea, in keeping with previous chemical tracer analyses. A rough calculation reveals that the amount of halocline water formed in this way in the Barents Sea and Fram Strait is 30-50% of that formed by ice growth in eastern Arctic polynyas. 1. Introduction Where does the "cold halocline" layer of the Arctic Ocean come from, and how does it form? In their seminal paper on the subject, Aagaard et al. [1981] review the previous observations and theories concerning this cold, salty layer. Figure 1 [Aagaard et al., 1981, Figure 4] shows the typical temperature-salinity structure of the upper 400 m. A cold, fresh surface mixed layer lies above a layer in which the temperature remains near the freezing point while the salinity increases with depth. At about 100 m depth the temperature also begins to increase, up to a maximum within the core of the warm, salty Atlantic layer. The "kink" in the temperature-salinity curves at about 100 m marks the core of the cold halocline layer. If this layer formed by simple vertical mixing between water in the surface and Atlantic layers, it would lie on a straight line between the two in Figure 1. Since it does not, it is a distinct water mass that must advect into the Arctic Ocean from its periphery. Further, were it not for the presence of this layer, the heat in the Atlantic layer would be able to mix up into the surface waters and melt the overlying sea ice. As it is, turbulence in the mixed layer only entrains cold, salty water from below and therefore has little effect on the heat budget of the surface layer. Aagaard et al. [ 1981] discuss two mechanisms for generating cold halocline water, shown in Figure 1 by two arrows. The first is salinization of relatively fresh shelf waters during winter, as ice freezes in the seasonal ice pack and salt is rejected back into the water. The analysis by Aagaard et al. [1981] indicated that the northern Barents Sea is a particularly likely source of halocline water formed by this mechanism. Coastal polynyas of the Arctic Ocean are another potential source, given the high rate of ice formation that occurs ...

Leads in pack ice have long been considered important to the thermodynamics of the polar regions. A winter lead affects the ocean around it because it is a density source. As the surface freezes, salt is rejected and forms more dense water which sinks under the lead. This sets up a circulation with freshwater flowing in from the sides near the surface and dense water flowing away from the lead at the base of the mixed layer. If the mixed layer is fully turbulent, this pattern may not occur; rather, the salt rejected at the surface may simply mix into the surface boundary layer. In either event the instability produced at the surface of leads is the primary source of unstable buoyancy flux and, as such, exerts a strong influence on the mixed layer. Here as many as possible of the disparate and almost anecdotal observations of lead oceanography are assembled and combined with theoretical arguments to predict the form and scale of oceanographic disturbances caused by winter leads. The experimental data suggest the velocity disturbances associated with lead convection are about 1–5 cm s−1. These appear as jets near the surface and the base of the mixed layer when ice velocities across the lead are less than about 5 cm s−1. The salinity disturbances are about 0.01 to 0.05 psu. Scaling arguments suggest that the geostrophic currents set up by the lead density disturbances are also of the order of 1–5 cm s−1. The disturbances are most obvious when freezing is rapid and ice velocity is low because the salinity and velocity disturbances in the upper ocean are not smeared out by turbulence. In this vein, lead convection may be characterized at one extreme as free convection in which the density disturbance forces the circulation. At the other extreme, lead convection may be characterized as forced convection in which the density disturbance is mixed rapidly by boundary layer turbulence. The lead number Lo, which is the ratio of the pressure term to the turbulence term in the momentum equation, and the turbulent lead number Lot, which is the ratio of buoyant production to shear production in the turbulent kinetic energy equation, define the boundary between the free and forced regimes. For Lo and Lot less than one, both the large‐scale circulation and the turbulence are forced by surface stress. For Lo and Lot greater than one, both the large‐scale circulation and the turbulence are forced by the buoyancy flux. The magnitudes of velocity and salinity disturbances from a model developed elsewhere, suitable to free convection, agree with what few observations we have. The results of a forced convection model, developed here, suggest salinity disturbances of the order of 0.01–0.02 practical salinity units, with the maximum occurring at the surface of the lead and decreasing substantially below 5–10 m. This unstable gradient is a unique characteristic of lead convection. Though the salinity disturbances may be small when ice velocities are large, the buoyancy flux in leads has a major effect on the boundary layer turbulence.

No abstract

Hierarchy implies that the study of sea ice can be divided into analysis of subsets of processes based on scale and their interaction with adjacent scales. We apply these concepts to regional sea ice dynamics. The apparent self‐similar property of ice floes seen in aircraft or satellite images argues for an aggregate nature of sea ice, that viscouslike regional behavior arises from discrete floe interactions. However, for some regions and some times, characteristic behavior, where lead patterns seen in basin‐wide advanced very high resolution radiometer images appear to be related to coastal orientation hundreds of kilometers away, suggests that small regional scale processes O(10 km) and discontinuities in the velocity or stress state along boundaries can affect the larger‐scale sea ice distribution and dynamics O(500 km). Thus sea ice displays both aggregate type behavior and discontinuous type behavior based on the history of forcing and shape of the enclosing basin. The appropriate matching of atmospheric processes to sea ice processes in air‐ice interaction is through the sea ice deformation field rather than the response of ice velocity to the local wind. This is because atmospheric forcing and sea ice deformation have matching energetic scales at several hundred kilometers and timescales of days. An example of northerly winds during the April 1992 Arctic Leads Experiment period suggests discontinuous type behavior upwind of the Alaska coast followed by a general opening behavior with easterly winds. There appear to be natural scale divisions between climate scale sea ice processes of O(100–300 km) which resolve aggregate behavior, regional scale O(10–50 km) which is necessary to resolve observed shearing behavior, and the floe scale O(1 km). Because the climate scale is two levels removed from the floe scale, care must be exercised in using ice properties from the floe scale in climate scale models; ice strength is an example of such a scale dependent parameter.

No abstract

No abstract

Three examples of inter-agency cooperation utilizing current generation, individual Autonomous Underwater Vehicles (AUVs) are described consistent with recent recommendations of the U.S. Commission on Ocean Policy. The first steps in transforming individual AUVs into adaptive, networked systems are underway. To realize an affordable and deployable system, a network-class AUV must be designed with cost–size constraints not necessarily applied in developing solo AUVs. Vehicle types are suggested based on function and ocean operating regime: surface layer, interior and bottom layer. Implications for platform, navigation and control subsystems are explored and practical formulations for autonomy and intelligence are postulated for comparing performance and judging behavior. Laws and conventions governing intelligent maritime navigation are reviewed and an autonomous controller with conventional collision avoidance behavior is described. Network-class cost constraints can be achieved through economies of scale. Productivity and efficiency in AUV manufacturing will increase if constructive competition is maintained. Constructive strategies include interface and operating standards. Professional societies and industry trade groups have a leadership role to play in establishing public, open standards.