An ice pump is a heat engine, driven by the change of freezing point with pressure, which will melt ice at depth in the ocean and deposit it at a shallower location: it is self‐starting. Calculations of the maximum magnitude of this effect are made which show good agreement with field data available for sea and lake ice. The discussion is applied to the general case of a moving pack ice sheet with a well‐mixed surface layer and to floating ice shelves. The rate of melt from an 11‐m‐deep pressure ridge keel due to ice pumping is estimated as 26 cm/year, and that from the front of the Ross Ice Shelf at McMurdo Sound, Antarctica is estimated as 5 m/year for the level of water movement noted in the authors' field observations. Far from the ice front, pumping between shelf areas of different thickness will still occur, with tidal motion providing the necessary water exchange, but its magnitude is now limited by the ability to remove the potentially stable layer of melt water out of the system. It is important to realize that the pumping does not depend on the availability of sensible heat in the water column and its effects are additional to any melting caused by the advection of warmer water to the ice‐water interface.

No abstract

No abstract

Potential temperature (θ) and salinity (S) data obtained along the perimeter of the southern Canadian Basin north of the East Siberian Sea in 1993 aboard the CCGS Henry Larsen show higher temperatures in waters of Atlantic origin than in available climatological data for the Canadian Basin. In particular, a front is observed near the Mendeleyev Ridge which separates the cooler Atlantic waters of the Canada Basin from the warmer Atlantic waters observed in the Makarov Basin. The front is further characterized by a change in the θ/S slope of Arctic thermocline water, and by thermohaline intrusions (θ and S reversals) within the Atlantic layer. The idea that this warm variety of Atlantic water has come recently from the Eurasian Basin is supported by its higher level of the tracer CFC‐11.

No abstract

No abstract

No abstract

Salinity data used to trace water movement or compute density are normally derived from measurements of chlorinity or electrical conductivity, temperature, and pressure. The latter technique has a precision about 1 order of magnitude greater than that of a typical chlorinity titration, but both are sensitive, in different ways, to variations in the ionic ratios of seawater. Present definitions of salinity are also ion dependent, causing significant variations in the salinity‐density relationship which cannot be simply expressed. In order to obtain density to an accuracy commensurate with the available precision it is best to define salinity in relation to a water mass of known ionic content so that a density correction to be applied to other water masses may be expressed as variations from a fixed standard. These corrections then appear in the form of simple additive constants for most waters, and where density difference is the important parameter, no correction is necessary within a specific water mass. The new salinity definition is based on dilution by weight of a conductivity ratio labeled standard seawater. It would be invariant under compositional variations and in accord with the proposed new equation of state (Grasshoff, 1976). It is conservative within acceptable limits, would provide a ‘practical salinity scale’ for use by oceanographers of all levels of sophistication, and would greatly facilitate data comparisons between institutions. The present variety of computational procedures for in situ data reduction would be replaced by one set of definitive equations that would not be subject to change as the precision of physical or chemical measurement improved. A great part of the data base necessary to write these equations exists, and the remainder should be available by 1978.