Clemens, Steven C; Farrell, John W; Gromet, L Peter (1993): (Table 1) Strontium isotope ratios of ODP Hole 121-758A [dataset]. PANGAEA, https://doi.org/10.1594/PANGAEA.769903,

---

Supplement to: Clemens, SC et al. (1993): Synchronous changes in seawater strontium isotope composition and global climate. Nature, 363(6430), 607-610, https://doi.org/10.1038/363607a0

The 87Sr/86Sr ratio of sea water has increased gradually over the past 40 Myr, suggesting a concomitant increase in global chemical weathering rates (Raymo et al., 1988, doi:10.1130/0091-7613(1988)016<0649:IOLCMB>2.3.CO;2; Capo and DePaolo, 1990, doi:10.1126/science.249.4964.51; Hodell et al., 1990, doi:10.1016/0168-9622(90)90011-Z; Raymo and Ruddiman, 1992, doi:10.1038/359117a0; Caldeira, 1992, doi:10.1038/357578a0; Palmer and Edmons, 1992, doi:10.1016/0016-7037(92)90332-D). Recently, Dia et al. (1992, doi:10.1038/356786a0) analysed a 250-kyr 87Sr/86Sr record, and found superimposed on this gradual increase higher-frequency 87Sr/86Sr variations which appeared to follow a 100-kyr cycle; this periodicity corresponds to one of the prominent cycles in the Earth's orbital parameters, which are known to modulate the patterns of solar insolation and hence climate (Berger, 1978, doi:10.1016/0033-5894(78)90064-9; 1989, doi:10.1016/1040-6182(89)90016-5; Imbrie et al., 1992, doi:10.1029/92PA02253). The resolution of this record was, however, insufficient to establish the phase relationship between the 87Sr/86Sr variations and global climate cycles. Here we present a high-resolution seawater 87Sr/86Sr record spanning the past 450 kyr. We find that maxima and minima in 87Sr/86Sr coincide with minima and maxima, respectively, in continental ice volume (from the SPECMAP oxygen isotope record (Imbrie et al., 1984)), apparently suggesting that there was less chemical weathering in arid glacial periods than in the more humid interglacials. During glacial-interglacial transitions, however, seawater 87Sr/86Sr changes at a rate of ~1 p.p.m./kyr, approximately three times that evaluated by Dia et al. (1992, doi:10.1038/356786a0). Mass-balance calculations illustrate that simple changes in modern chemical weathering regimes cannot fully account for such rapid changes, suggesting that we need to revise current ideas about strontium reservoirs and the mechanisms for exchange between them.

Strontium isotopes were measured on the acetic acid-soluble portions of five or ten hand-picked and ultrasonically cleaned tests of the planktonic foraminifera Globorotalia menardii (>425 µm). Samples were analysed for 87Sr/86Sr by thermal ionization mass spectrometry with a Finnigan-MAT 261 mass spectrometer using static multiple collection with simultaneous collection of mass 85 to monitor for the presence of Rb. The filament current was raised over a period of 45 min to yield a stable 88Sr signal of ~4 * 10**11 A. We did not detect a statistically resolvable signal at mass 85, and any possible corrections are of the order of 2-3 * 1**-6. Results were normalized for mass fractionation to a 86Sr/87Sr ratio of 0.1194 according to an exponential fractionation law. The site 758 87Sr/86Sr ratio measurements were made between May and October 1992 during which time one or more aliquots of the standard NBS-987 were analysed with each spectrometer run (6-12 samples). The external error associated with NBS-987 replicate analyses is ±0.000012 (1 sigma, N=112). We report results relative to an NBS-987 value of 0.710150. A typical analysis consisted of 14 blocks of 15 scans lasting 32 seconds each (210 ratio counts). To estimate sample precision, which accounts for sample inhomogeneity and sample processing, we made 91 replicate analyses from raw sample (168 total analyses). Each replicate was passed through the entire chemical procedure. The resulting standard deviation (±0.000011; 1 sigma) agrees well with the value calculated from standards. The ± 1 sigma reported with each sample is calculated as the long-term external error (±0.000012) divided by n**0.5 where n is the number of replicate analyses from raw sample (typically 2). Two full procedural blanks run during data collection contained 32 and 96 pg of Sr.