Annual sea-air CO2 fluxes in the Bering Sea: Insights from new autumn and winter observations of a seasonally ice-covered continental shelf

High-resolution data collected from several programs have greatly increased the spatiotemporal resolution of pCO2(sw) data in the Bering Sea, and provided the first autumn and winter observations. Using data from 2008 to 2012, monthly climatologies of sea-air CO2 fluxes for the Bering Sea shelf area from April to December were calculated, and contributions of physical and biological processes to observed monthly sea-air pCO2 gradients (ΔpCO2) were investigated. Net efflux of CO2 was observed during November, December, and April, despite the impact of sea surface cooling on ΔpCO2. Although the Bering Sea was believed to be a moderate to strong atmospheric CO2 sink, we found that autumn and winter CO2 effluxes balanced 65% of spring and summer CO2 uptake. Ice cover reduced sea-air CO2 fluxes in December, April, and May. Our estimate for ice-cover corrected fluxes suggests the mechanical inhibition of CO2 flux by seaice cover has only a small impact on the annual scale ((\u3c2%). An important data gap still exists for January to March, the period of peak ice cover and the highest expected retardation of the fluxes. By interpolating between December and April using assumptions of the described autumn and winter conditions, we estimate the Bering Sea shelf area is an annual CO2 sink of 6.8 Tg C yr21. With changing climate, we expect warming sea surface temperatures, reduced ice cover, and greater wind speeds with enhanced gas exchange to decrease the size of this CO2 sink by augmenting conditions favorable for greater wintertime outgassing