Auxiliary Material for Paper 2011GB004141 Southern Ocean control of silicon stable isotope distribution in the deep Atlantic Ocean Gregory F. de Souza and Ben C. Reynolds Institute of Geochemistry and Petrology, ETH Zurich, Zurich, Switzerland Jorg Rickli Institute of Geochemistry and Petrology, ETH Zurich, Zurich, Switzerland Now at Bristol Isotope Group, University of Bristol, Bristol, UK Martin Frank GEOMAR, Helmholtz Centre for Ocean Research Kiel, Kiel, Germany Mak A. Saito Woods Hole Oceanographical Institution, Woods Hole, Massachusetts, USA Loes J. A. Gerringa Royal Netherlands Institute for Sea Research, Texel, Netherlands Bernard Bourdon Institute of Geochemistry and Petrology, ETH Zurich, Zurich, Switzerland Ecole Normale Superieure de Lyon and CNRS, Lyon, France de Souza, G. F., B. C. Reynolds, J. Rickli, M. Frank, M. A. Saito, L. J. A. Gerringa, and B. Bourdon (2012), Southern Ocean control of silicon stable isotope distribution in the deep Atlantic Ocean, Global Biogeochem. Cycles, 26, GB2035, doi:10.1029/2011GB004141. Introduction: This readme file contains information regarding the Auxiliary Material submitted to GBC. The Auxiliary Material consists of the Auxiliary Table 1 (2011GB004141-auxtable1.txt and 2011GB004141-auxtable1.xls) as well as a single PDF file (2011GB004141-auxiliarymaterial.pdf) including additional information that could not be included within the manuscript. Details are given in the following. 1. 2011gb004141-ts01.txt, 2011gb004141-ts01.xls Table S1. This table contains the primary dataset of delta(30)Si values presented in this manuscript, along with ancillary hydrographic and chemical data. It is provided both in tab-delimited ASCII as well as formatted XLS (MS Excel) formats. The authors recommend referring to the XLS file. The delta(30)Si data presented in this table are the basis of Figs. 2, 3 and 4 of the manuscript, as well as Figs. S1-S4 of the Auxiliary Material (see 3.1 below). 1.1 Column "Cruise", expedition designation. 1.2 Column "Station", sampling station designation. 1.3 Column "Depth", meters, sampling depth. 1.4 Column "Theta", degrees Celsius, potential temperature. 1.5 Column "Salinity", practical salinity units, salinity. 1.6 Column "sigma_theta", kg/m^3, potential density anomaly. 1.7 Column "[PO4]", micromoles/kg, phosphate concentration. 1.8 Column "[NO3]", micromoles/kg, nitrate concentration. 1.9 Column "[Si]", micromoles/kg, silicon concentration. 1.10 Column "delta(30)Si", permil, silicon stable isotope composition vs. NBS 28 as defined in Section 2.2 of the manuscript. 1.11 Column "2sigma_SEM (external)", permil, external 2 standard errors of the mean uncertainty on delta(30)Si data (see Section 2.2 of the manuscript). 1.12 Column "n", unitless, number of complete analytical replicates for isotopic analyses 1.13 Column "PO4*", micromoles/kg, the quasi-conservative tracer PO4* as defined in Auxiliary Material B (see 3.3 below) 1.14 Column "Forced PO4*", micromoles/kg, identifies two samples with PO4* values below 0.76 micromoles/kg that were arbitrarily corrected to this value, as detailed in Auxiliary Material B and C (see 3.3 and 3.4 below). This column is presented as parentheses in Column "PO4*" in the XLS version of the table. 1.15 Column "Deltadelta(30)Si_msr-mix", permil, the deviation of measured delta(30)Si values from those expected from pure binary mixing, as given by the equation in Section 4.1 of the manuscript. 1.16 Column "Forced Deltadelta(30)Si_msr-mix", permil, Deltadelta(30)Si_msr-mix values of the two samples with PO4* below 0.76 micromoles/kg after arbitrary correction to 0.76 micromoles/kg (see Column "Forced PO4*"). This column is presented as parentheses in Column "Deltadelta(30)Si_msr-mix" in the XLS version of the table. 2. 2011gb004141-txts01.pdf Text S1. The following subsections provide information on the auxiliary material provided within the PDF file. 2.1 Auxiliary Figures The first section of the Auxiliary Material includes depth profiles of delta(30)Si and [Si] for all stations included in the study, grouped by geographical location. Captions are as below: Fig. S1: High northern latitude stations (expedition 64PE319) Fig. S2: Labrador Sea and temperate eastern (ANT23/1) and western (64PE319) Atlantic stations Fig. S3: Eastern Atlantic stations from expedition ANTXXIII/1, with Drake Passage profile for comparison (note different d30Si scale from previous figures) Fig. S4: Eastern (Sta 13) and western (Sta 3) South Atlantic stations (expedition KN192-5) 2.2 Section A: Standard addition test of accuracy This section provides information regarding the standard addition tests used to verify accuracy of the d30Si analyses. A figure is included that shows the results of two such tests. Fig. S5: Plots showing the linear variation of delta(30)Si values of standard-sample mixtures with the fraction of standard in the mixture, f_STD. Filled circles at the extremes of the x-axis represent the pure sample and standard (at x = 0 and x = 1 respectively); mixtures are represented as open error ellipses. Linear regressions utilize only mixture data and are calculated in Isoplot (K. R. Ludwig, Berkeley Geochronological Center) using the algorithm of York [1968]. Mean square weighted deviation (MSWD) values below unity would seem to indicate that the errors on the datapoints (2SD external) are slightly overestimated. 2.3 Section B: Endmember contribution calculations This section details the method followed for calculation of the proportion of Si in a sample contributed by the North Atlantic endmember (f_NA), as discussed in Section 4.1 of the manuscript. It provides the equations used, together with a discussion of the choice of endmember characteristics and of the errors associated with the endmember contribution calculations. 2.4 Section C: Endmember contribution calculations: PO4* correction This section details the correction of cruise data that was necessary to produce consistency between datasets from expedition 64PE319 and the TTO-NAS programme. Numerous figures are included to allow the reader to follow the rationale for this correction. Fig. S6: Temperature-salinity comparisons of the TTO-NAS and 64PE319 stations used to correct PO4* values. Fig. S7: Comparison of potential density (sigma_theta) profiles for the TTO-NAS and 64PE319 stations. Fig. S8: Comparison of depth profiles of PO4* from TTO-NAS and 64PE319 stations. Fig. S9: Comparison of PO4*–potential density relationships between TTO-NAS and 64PE319 stations. Fig. S10: PO4* values for samples corrected by addition of 0.05 mmol/kg to calculated values for 64PE319 samples (>= 2000 m water depth), in comparison to TTO data. Note that PO4* values from Sta 2, from which the raw data show the strongest offset to TTO data, continue to exhibit a discrepancy to TTO data even after correction, while the other stations show good consistency. 64PE319 Sta 3 is also shown in upper right panel for comparison, although no TTO-NAS station is sufficiently close to it to provide good control. 2.5 Section D: Box model This section included details of the box model architecture, handling of isotope tracers and parameter values that were not included in the main manuscript's text or figures, as well as a discussion of model parameter value choices and their effects on model results, exemplarily illustrated by a figure. A table lists the key model parameters. Fig. S11: Model results for two extreme sensitivity test configurations with surface opal productivity rates decreased to 50% (panel a) or increased to 200% (panel b) of the value in standard configuration (Table S1). All other parameter values remain unchanged (although the constraint of 1 microM in the low-latitude surface continues to apply). Note the change in [Si] gradients between boxes s and p, as well as between boxes ds and dn, and also the change in the magnitude of the deep d30Si gradient. In model a, total export is 75 Tmol Si yr-1 with the Southern Ocean contributing 30% of this. In model b, total export is 107 Tmol Si yr-1, with a Southern Ocean contribution of 66%. Table S1: Parameters of the 8-box model utilized in this study.