Fall case studies of three-dimensional circulation, plankton, and benthos models explored the consequences of interannual changes in ice cover and water motion on carbon/nitrogen cycling by the end of September within the Chukchi/Beaufort Seas. The coupled model scenarios were those of reduced (greater) northward flow, colder (warmer) temperatures, and more (less) extensive ice cover over the preceding ~60 days of August and September during the negative (positive), anticyclonic (cyclonic) phase of the Arctic Oscillation in 1980 (1989). On the inner Chukchi shelf, stronger flows in 1989 advected nitrate and silicate stocks of Pacific origin ~130 km farther northwest toward Wrangel Island than in 1980. Yet an increase of the total net photosynthesis by the diatom-dominated phytoplankton community over both shelves in 1989 was mainly the result of less ice cover of the cyclonic period, with a concomittant increase of POC influxes of phytodetritus and fecal pellets to the sediments. In terms of present shelf export, the model's separate pools of ~65 umol DOC kg-1 and 1 ug chl l-1, or ~4 umol POC kg-1, at a depth of 60 m above the 2000-m isobath of the Beaufort Sea in September 1989, matched the sum of ~70 umol TOC kg--1 sampled there by submarine in September 1997. Accordingly, most of the simulated Chukchi shelf was a weak sink of atmospheric CO2 in both September 1980 and 1989, reflecting a net fall export of particulate and dissolved debris. Within the cyclonic case of strong flows in 1989, a surface pCO2 of 248 uatm was also simulated in September at 155¿W on the Beaufort shelf, where ~250 uatm was measured there in September 2000. Here, farther away from the Pacific source of nutrients for enhanced photosynthesis, the model's estimate of surface sea water fugacity in a weaker flow regime was only 375 uatm of pCO2 at the same location in September 1980, when typically outgassing would have instead prevailed, despite increasing atmospheric pCO2 values, i.e., 356 to 362 uatm of pCO2 were found in Arctic air during 1998 to 2000. Although the northward model transports of water, gases, nutrients, and particulate matter were less during 1980 than in 1989, the simulated relict nutrient fields, left behind by the model's phytoplankton on the outer shelf of the Chukchi Sea, were greater then, as a consequence of their smaller rates of light-limited utilization during less photosynthesis and draw down of CO2. The more important factors, effecting fluxes of CO2 here, may thus be the biological ones of photosynthesis and respiration, rather than the physical ones of temperature and advection. Since the colonial prymnesiophytes of the model also grew better atlow light intensities than the diatoms or microflagellates, Phaeocystis won, when all other constraints were similar, in regions of the ice-covered upper euphotic zone, where total stocks were less than 0.5 ug chl l-1, as observed in August 2000. However, increased model fidelity, compared with submarine observations of nutrients along the shelf break of the western Canadian Basin, will require appropriate specification of the persistence of winter initial conditions in Bering Strait. Future reductions of ice cover over deeper regions of these Arctic shelves may result in greater utilization of the untapped winter nutrients by these shade-adapted prymnesiophytes, with less energy transfer to higher trophic levels of their food webs. Altered bases of shelf food webs, rather than increased northward transport of Pacific waters and their constituents, may then have greater consequences for both future element cycling and protein yield of these western Arctic ecosystems.