•The fMW in upper water column increased along pathway of the Pacific inflow.•228Ra/226Ra)A.R. correlates well with salinity and fraction of meteoric water.•Transit time of river water from Bering Shelf to Canada Basin was estimated as 10.4–17.6 years.Seawater samples for the measurements of 226Ra, 228Ra and stable oxygen isotope (δ18O) were collected from the Bering and Chukchi Seas in the summer of 2014. The fractions of meteoric water (fMW) and sea-ice melted water (fSIM) were estimated based on the mass balance of salinity and δ18O with a three end-member mixing model. Our results showed that the average fMW increased northward from the Bering Basin to the Canada Basin while the fSIM distributed homogeneously. The lowest fMW and 228Ra/226Ra)A.R. values were found in the upper Bering Basin with little terrestrial input. The highest fMW but low 228Ra/226Ra)A.R. appeared in the northern Chukchi Sea and the Canada Basin, ascribing to the current-driven accumulation of freshwater and its long residence time. More abundant sea-ice melted water was found on the pack-ice edge, indicating the trap of earlier melted waters by the ice pack. Based upon the linear relationships between 228Ra/226Ra)A.R. and fMW in the Bering Shelf and the Chukchi Shelf, the transit time for the Pacific inflow was constrained. The transit time of river water from the Bering Shelf to the Chukchi Shelf was estimated as 0.2–4.4 years with an average of 1.6 ± 1.5 years, while that from the Chukchi Shelf to the Canada Basin was 10.2–13.2 years with an average of 11.8 ± 1.1 years. The spatial variation of the transit time was mainly affected by the current intensity. Our study highlights the importance of in-depth evaluation for the subarctic-arctic exchange.

•Macromolecules and COM preferentially complexed with 234Th over 233Pa in seawater.•NOM affects their adsorption on inorganic nanoparticles through organic coatings.•NOM and nanoparticles synergistically affect the partitioning of 234Th and 233Pa.•A strong ‘colloidal concentration effect’ was observed.Laboratory experiments were carried out to examine the role of natural organic matter (NOM) in regulating the adsorption and fractionation of 234Th and 233Pa in seawater on inorganic nanoparticles (20 nm), including SiO2, CaCO3, Fe2O3, Al2O3 and TiO2. Model macromolecules and natural colloidal organic matter (COM) tested in the present study included dextran with different molecular weights, acid polysaccharides (APS), humic acids (HAs), protein (bovine serum albumin, BSA), and COM isolated from river water and seawater. APS and HAs had comparable affinity and higher partition coefficients (Kd) for 234Th, with logKd values close to 7, while proteins showed weaker binding strength with 234Th and had a logKd value of 5.75. Compared to 234Th, 233Pa showed lower affinity for different organic matters, with logKd values ranging from 4.41 on APS to 5.46 on 6-kDa-dextran. In general, macromolecules and COM preferentially complexed with 234Th over 233Pa, which resulted in a fractionation between 234Th and 233Pa on different NOMs with a fractionation factor (FTh/Pa) following the order of APS > HA ≥ Dextran > BSA. Results from binary-sorbent experiments (nanoparticles plus NOM) indicated that NOMs evidently affected the adsorption of 234Th and 233Pa on different inorganic nanoparticles likely through the formation of organic coatings. While SiO2 preferentially adsorbed 233Pa over 234Th, the presence of APS could compete the effective sorption sites and thus reduced the adsorption of 233Pa on SiO2 nanoparticles. Based on a two end-member mixing model (Li, 2005), coating extent seemed to depend on the relative abundance and chemical composition of organic matter and inorganic nanoparticles. In addition to organic composition, NOM concentrations also affected the partitioning of 234Th and 233Pa between seawater and particles, showing a strong ‘colloidal concentration effect’. Nevertheless, there was little change in logKd values of 234Th and 233Pa on dextran with a molecular weight between 6–500 kDa or 2–10 nm in size, indicating negligible NOM size effect from small colloids on the adsorption of Th and Pa in seawater.

Surface seawater was collected for 226Ra measurement in the North Pacific Subtropical Gyre from July to October, 1999 and October to December, 2003. Combined with the historical data reported for this sea area, a declined trend of surface 226Ra concentrations was observed since 1960s, indicating the ecosystem shift in response to global warming. On one side, the enhanced stratification of the upper water column resulting from global warming reduced the 226Ra input from the depth, on the other, the temporal increase of biological production resulting from the climate-related ecosystem structure change strengthened the 226Ra removal from the surface ocean. Both the physical and biological processes resulted in the decrease of surface 226Ra concentrations in the North Pacific Subtropical Gyre. The temporal trend of surface 226Ra concentrations was consistent with the trends of chlorophyll a, silicate, phosphate and primary production previously reported. This study provided 226Ra evidence for the ecosystem shift under global change.

•δ13CPOC in the inner Prydz Bay was enriched relative to that in open ocean waters.•A negative correlation between δ13CPOC and the fraction of meltwater estimated from 226Ra-salinity mass balance was first reported in the Antarctic waters.•Ice melting may affect surface ocean δ13CPOC by changing water column stability and the subsequent physical–biological coupling.The stable carbon isotope composition of particulate organic carbon (δ13CPOC) and naturally occurring long-lived radionuclide 226Ra (T1/2=1600 a) were applied to study the variations of upper ocean (<100 m) carbon dynamics in response to sea ice melting in Prydz Bay, East Antarctica during austral summer 2006. Surface δ13CPOC values ranged from −27.4‰ to −19.0‰ and generally decreased from inner bay (south of 67°S) toward the Antarctic Divergence. Surface water 226Ra activity concentration ranged from 0.92 to 2.09 Bq/m3 (average 1.65±0.32 Bq/m3, n=20) and increased toward the Antarctic Divergence, probably reflecting the influence of 226Ra-depleted meltwater and upwelled 226Ra-replete deep water. The fraction of meltwater, fi, was estimated from 226Ra activity concentration and salinity using a three-component (along with Antarctic Summer Surface Water, and Prydz Bay Deep Water) mixing model. Although the fraction of meltwater is relatively minor (1.6–11.9%, average 4.1±2.7%, n=20) for the surface waters (sampled at ~6 m), a positive correlation between surface δ13CPOC and fi (δ13CPOC=0.94×fi−28.44, n=20, r2=0.66, p<0.0001) was found, implying that sea ice melting may have contributed to elevated δ13CPOC values in the inner Prydz Bay compared to the open oceanic waters. This is the first time for a relationship between δ13CPOC and meltwater fraction to be reported in polar oceans to our knowledge. We propose that sea ice melting may have affected surface ocean δ13CPOC by enhancing water column stability and providing a more favorable light environment for phytoplankton photosynthesis, resulting in drawdown of seawater CO2 availability, likely reducing the magnitude of isotope fractionation during biological carbon fixation. Our results highlight the linkage of ice melting and δ13CPOC, providing insights into understanding the carbon cycling in the highly productive Antarctic waters.

Activity concentrations of dissolved and particulate 210Pb and 226Ra in the water column were measured in the eastern Chukchi Sea during summer 2003. 234Th/238U disequilibria were used to estimate the scavenging fluxes of 210Pb from the water column to the underlying sediments. Our results showed that concentrations of 210Pb and its distributions were mainly influenced by mixing processes of water masses and sediment resuspension. The residence times of 210Pb in the eastern Chukchi Sea ranged from 5 to 103 d. Short residence times were mostly observed at the shelf stations, indicating a more effective particle scavenging in the shelf region. A mass balance model was constructed to evaluate the contribution of lateral transport to 210Pb budget in the water column. The lateral transport fluxes of 210Pb ranged from 17 to 177 Bq/m2/a, comprising up to 63–94% of the total supply of 210Pb in the eastern Chukchi Sea. We hypothesize that the accumulative removal of 210Pb in the Pacific inflow waters during their transport across the Chukchi Sea and the import of 210Pb from sea ice rafted sediments are the two major lateral transport pathways for the import of 210Pb to the eastern Chukchi Sea. Our results highlight the importance of lateral transport processes to the geographical distribution of particle-reactive elements and their biogeochemical cycles in the Arctic Ocean.

The relation between the dissolved carbon dioxide (CO2) and the stable isotopic composition of particulate organic matter in the water column has not been well quantified, but this information could help provide a better understanding of carbon dynamics in a warmer Arctic Ocean. The stable carbon isotopic composition of suspended particulate organic carbon (δ13CPOC) in the surface waters of the western Arctic Ocean was measured during July–September 2003, to evaluate the spatial variability of δ13CPOC and its key controlling factors. Values of δ13CPOC fell within the range of −28.5‰ to −21.1‰, with an average of −24.5±2.3‰. The spatial variability of δ13CPOC showed a general decreasing trend from shallow waters in the continental shelf toward the deeper, colder waters in the basin. A negative correlation between δ13CPOC and the dissolved CO2 concentration in surface waters was observed, indicating that carbon isotopic fractionation during photosynthesis was largely dependent on the dissolved CO2 concentration. Compared to the solubility pump, biological processes may play a more important role in determining the distribution and variation of δ13CPOC in the western Arctic Ocean during summer. The coupled relationship between CO2 concentration and stable isotopic composition of particulate organic matter has the potential to be used for reconstruction of sea-surface CO2 changes in the past, provided a quantitative relationship of δ13C between POC and sediments can be established.