/. This stable isotope is most abundant in ocean water and least 

 abundant in fresh precipitation and runoff, and so mixtures 

 show intennediate values in proportion, just like salinity; and 

 2. the freezing process, which is important in increasing salinity 

 of the northern shelf waters, does not alter the isotope abundance. 



The correlations between salinity and oxygen isotope for 

 the water masses are plotted in Fig. 8 (oxygen isotope data 

 from Grebmeier et ai, 1990). Isotope values are plotted as 

 deviations from standard mean ocean water ('^Osmow)- so that 

 most ocean waters have values close to and freshwater is 

 <-20 ppt. In Fig. 8, samples are plotted in two ways: as 

 individual point correlations for samples from the Bering Slope 

 Current (triangles), Alaskan Coastal (open circles) and Siberian 

 Coastal (solid circles) water masses, and as envelopes 

 encompassing many values for the other three. 



All samples lie on or very close to a line from S = 35, 

 "*0 = (ocean water) and S = 0, '"O = -24.6 (freshwater), from 

 which we can conclude that all water masses are essentially 

 simple dilutions of the most saline Bering Slope Current water 

 by freshwater. The progression along the line is orderly. The 



v'8. 



S 0/SALINITY 



ANADYR CURRENT 



BERING SLOPE CURRENT 



GULF OF ANADYR 



•• SIBERIAN 



COASTAL WATER 



ALASKAN COASTAL 

 O 



31 32 33 34 



BOTTOM WATER SALINITY. 7oo 



Fig. 8. Correlation of '"O with salinity. Koriilev cruise, August 1988. All 

 water masses of the Northern Bering Sea Ecosystem are dilutions to 

 varying degrees of Bering Slope Current water by freshwater. Gulf of 

 Anadyr and Siberian Coastal waters are modifications of Bering Shelf 

 water through salinity enhancement due to freezing. Arrows indicate 

 direction of water mass modification. Data from Grebmeier. Cooper 

 & DeNiro, 1990. 



precursor water to the whole system from the Bering Slope 

 Current has both the highest salinities and abundances of 

 '"O: S ~ 33 to 33.2, '"O - -1.5. These values are, of course, 

 already slightly diluted from SMOW. In the ecosystem, the 

 first step in dilution is observed in the Anadyr Current, because 

 the current mixes to some extent with runoff (particularly the 

 Anadyr River) in its transit around the Gulf of Anadyr. |In 

 Fig. 8, the pathways of water mass modification are indicated 

 by arrows.] Furtherdilution of AnadyrCurrent water produces 

 the Bering Shelf water inass, ubiquitous to the whole northern 

 shelf. Not all AnadyrCurrent water transits Anadyr Strait, but 

 some is detlected to the south of St. Lawrence Island, where it 

 meets and mixes with fresher waters from the Alaskan side of 

 the system, forming this water mass with slightly reduced 

 salinities and '"O. Samples from the Alaskan Coastal water 



mass show much greater and more variable dilutions because 

 of proximity to the high runoff along the eastern side of the 

 system. They also do not follow the dilution curve as closely 

 because selected areas are subject to strong local freezing and 

 brine rejection, for example within Norton Sound ( see Muench 

 et ill.. 1981 ). The value at -3.3/31.5 is a good case in point. 



The two secondary water masses of the system. Gulf of 

 Anadyr and Siberian Coastal, are both created from Bering 

 Shelf water through salinity enhancement by freezing. The 

 polynya south of St. Lawrence Island, as discussed , is the focal 

 point for the salinity enhancement which turns Bering Shelf 

 water into Gulf of Anadyr water ; the overwinter freezing 

 increases salinities by about 0.5 but without changing "O. 



The Siberian Coastal water mass is apparently created in 

 the same way. Bering Shelf water travels throughout the 

 system, well north into the Chukchi Sea, without appreciable 

 change in S. The whole system evidences very little lateral 

 diffusion and exchange between water masses, and the Bering 

 Shelf water, sandwiched in the middle, is effectively isolated 

 from runoff and hence dilution from both Alaska and Siberia. 

 In the Chukchi, vigorous freezing in certain areas in winter 

 causes substantial increases in S values without modifying '*0, 

 and this water is recirculated the following year as part of the 

 Siberian Coastal Current (see discussion below). 



Flow Field 



The Anadyr Current, the branch of the Bering Slope 

 Current that enters the Gulf of Anadyr near Cape Navarin and 

 continuously supplies the nutrients to fuel the ecosystem, is a 

 topographic boundary current of the eastern Bering Sea Shelf; 

 it is also, coincidentally, located along the western boundary of 

 the shelf. This was convincingly demonstrated by Kinder e/ al. 

 ( 1986) who employed both laboratory models and numerical 

 simulations, achieving results in very close agreement with 

 what we know of the Anadyr Current. 



The basic driving force is the sink for Bering Sea water 

 imposed by the northward flow through Bering Strait — that is, 

 the pressure head created by a -0.5 m height difference 

 between the Bering Sea and the Arctic Ocean (Stigebrandt, 

 1984). Thus, Bering Sea water must move northward across a 

 shoaling topography. In this situation, the topographic gradient, 

 f / h ~ 5 X 10 '' cm ' s ', is more than an order greater than the 

 variation of Coriolis parameter, (3 ~ 1 x 10" cm' s '. The 

 across-shelf flow is concentrated as a current along the lefthand 

 boundary facing upslope (Fig. 9). Notice in the simulations 

 that regardless of whether or not flow conditions are imposed 

 along the Bering Sea slope, the cross-shelf flow still forms the 

 same western boundary current on the shelL The numerical 

 simulations indicated a current width of 50 km and speeds of 

 10-20 cm s ', both in excellent agreement with available data 

 on the real current. Of course, within the Gulf of Anadyr, the 

 flow, being strongly steered along isobaths, actually circulates 

 clockwise around the gulf (cf. Fig. 1 ). 



Variability in flow of the Anadyr current is unknown. It 

 seems probable, however, that it is a much steadier flow than 

 those through Anadyr and Bering Straits. The large variability 

 in the latter flows, predominantly at periods of a day to a week. 



21 



