The scavenging of Th-230, Pa-231 and Be-10 was studied in the Pacific sector of the Southern Ocean along 170 degreesW using measurements from surface sediments, time-series sediment traps, and the water column. All sediment traps collected an annual flux of Pa-231 greater than the integrated rate of production by U-235-decay (up to 5.5 times greater) and a flux of Be-10 greater than the global-average deposition rate of Be-10 (up to 3.4 times greater). Fluxes of Th-230 were on average close to the production rate in the overlying water column. These large fluxes of Pa-231 and (10) Be, and high unsupported Pa-231/Th-231 and Be-10/Th-230 ratios in the sediments, are not associated with depletion of nuclides in the water column, and therefore are not the result of the classical boundary scavenging mechanism. We estimate that of the Pa-231 and Th-230 advected into the Southern Ocean as part of the large-scale overturning circulation, only those nuclides associated with the "upper limb" of this circulation, i.e. those that pass through the surface as part of the wind-driven Ekman flow, are scavenged efficiently. The majority of the nuclides advected into the Southern Ocean and associated with the "bottom water limb" are not scavenged to the sediments of the Southern Ocean, but are returned northward.Throughout the year, the unsupported Pa-231/Th-230 and Be-10/Th-230 ratios of material reaching sediment traps at similar to 1000 m at a given site increase with increasing particle flux. This behavior is explained by a conceptual model in which Th, Pa and Be are scavenged reversibly by fine-grained suspended particles at all depths, while Pa and Be are in addition scavenged strongly by diatom-rich aggregates in surface waters. Spatial variability in the annually averaged unsupported Pa-231/Th-230 and Be-10/Th-230 ratios of sinking particulate matter reflects primarily the variability in particle composition, and in the ratio and quantity of nuclides upwelled to the mixed layer, rather than variability in particle flux. (C) 2003 Elsevier Science Ltd. All rights reserved.
662AGTimes Cited:12Cited References Count:53