It has been proposed (Pegram et al., 1992) that osmium isotope variations inferred for sea water over the Cenozoic have their origin in changes in the relative supply of radiogenic Os from the continents (187Os/188Os = 1.2-3.6) and primitive Os from micrometeoritic or mantle material (187Os/188Os = .12035). The aim of this work was to reconstruct global riverine inputs to the oceans through the Pleistocene using changes in the osmium isotopic composition of sea water as an indicator of variations in continental chemical erosion intensity.
We present new data confirming a previously-reported (Oxburgh and Broecker, 1994) 187Os/188Os shift of 0.97 to 1.02 over the past 10,000 years in sediments from two cores from the East Pacific Rise. A similar shift at the previous termination (boundary between isotopic stages 5 and 6) is also revealed. High, 'interglacial' ratios of 1.02-1.04 appear to characterise much of the glacial times, with a relatively brief and rapid excursion to low values (of about 0.97) occurring at the end of the glacial periods.
The osmium in these cores is almost entirely hydrogenous in origin so that osmium derived from terrestrial or cosmic dust makes up a very small fraction of the total, and changes in the fluxes of these components are unlikely to be the cause of the observed shift. In addition, a shift of the same magnitude is observed in both cores despite accumulation rates that differ by a factor of two. Were changes in dust flux responsible for the isotopic excursions, the shifts should be more marked in the low accumulation rate core. A local hydrothermal origin for the signal is not supported by relatively constant accumulation rates of other hydrothermally-derived metals over the same interval. Thus we deem changes in sea water composition to be the most likely source of the isotopic deviations.
The homogeneity of osmium isotopes in the modern (and past) oceans remains to be irrefutably established, placing an important caveat on the interpretation of the data. However, if the signal observed in these two cores is global in nature, it most likely reflects changes in the delivery of radiogenic osmium from the continents. A shift in the osmium isotopic composition of sea water of this magnitude may be explained by a change of 10-30% in the riverine flux of radiogenic osmium, consistent with decreased global weathering rates during late glacial times. Alternatively, it may reflect a change in the average isotopic composition of riverine osmium due to a decrease in the weathering of radiogenic continental shield rocks as they are covered by the advancing Laurentide ice sheet. Osmium isotopic analysis of nodules recovered from the Baltic Sea (Peucker-Ehrenbrink and Ravizza, 1995) (187Os/188Os = 3.7 - 7.8) support the hypothesis that the Baltic shield provides a source of radiogenic osmium to the oceans. Information on the relative fluxes and isotopic compositions of sea water osmium sources is required in order to defnitively resolve the origin of the Pleistocene signal.
If the osmium isotopic shift observed in these two cores indeed reflects a change in the mean composition of sea water, it clearly provides no support for increased weathering of continental rocks during glacial times. Thus the reduced concentrations of CO2 in the glacial atmosphere as recorded by bubbles trapped in glacial ice do not originate in increase riverine alkalinity fluxes during glacial periods.
Oxburgh, R. & Broecker, W.S., EOS 75, 362 (1994).
Pegram, W.J., Krishnaswami, S., Ravizza, G. & Turekian, K.K., Earth Planet Sci. Lett. 111, 569-576 (1992).
Peucker-Ehrenbrink, B. & Ravizza, G., EOS 76(46), 319 (1995).