Supporting Online Material for

Transcription

1 Supporting Online Material for Synchronizing Rock Clocks of Earth History K. F. Kuiper, A. Deino, F. J. Hilgen, W. Krijgsman, P. R. Renne, J. R. Wijbrans This PDF file includes: Materials and Methods SOM Text Figs. S1 to S4 Tables S1 to S4 References Published 25 April 2008, Science 320, 500 (2008) DOI: /science

2 Supporting Online Material The supporting online material includes: Material and Methods with details on data treatment both at BGC and the VU. All relevant analytical information (tables S1, S2, fig. S1). Age and error equations for astronomically calibrated standard. 40 Ar/ ages for the Matuyama- Bruhnes geomagnetic polarity reversal (table S3). Details of revised astronomical tuning of the K-T boundary (figs. S2-4). Comments on existing controversy of the relation between Chixculub impact crater and K-T boundary event. Details on 40 Ar/ data of the K-T boundary (table S4). References.

3 Material and Methods Volcanic levels intercalated in cyclic sedimentary sequences of marl and diatomites were sampled during joint field campaigns. Mineral separations and 40 Ar/ analyses were performed independently both at the Vrije Universiteit Amsterdam (VU) and the Berkeley Geochronology Center (BGC). All samples were irradiated in the OSU TRIGA reactor CLICIT facility in several batches ranging from 7-20 hours irradiation time. At BGC single crystals of sanidine were either fused or incrementally heated. J-values are calculated by a planar regression over the irradiations discs. At the VU single crystals of sanidine were fused. J values are calculated by a 2 nd order polynomial regression over the vial height. The neutron flux monitor FCs of ± 0.28 Ma (S1) is used in all irradiations apart from VU37, which used 85G003 of ± 0.28 Ma (S1). Residuals calculated for J-values are on the order of 0.1% both at BGC and the VU, in practice a slightly higher uncertainty (± % 1σ uncertainty in J) is arbitrarily used in the age calculations. Decay constants of (S2) are used in age calculations. Corrections factors for neutron interference reactions are 2.65 ± for ( 36 Ar/ 37 Ar) Ca, 6.95 ± for ( / 37 Ar) Ca, 1.22 ± for ( 38 Ar/ ) K and 7.3 ± for ( 40 Ar/ ) K at BGC and 2.64 ± for ( 36 Ar/ 37 Ar) Ca, 6.73 ± for ( / 37 Ar) Ca, ± for ( 38 Ar/ ) K and 8.6 ± for ( 40 Ar/ ) K at the VU (1 sigma errors). Note that all VU data have been recalculated with BGC correction factors resulting in negligible an age difference of 0.007% or less. Irradiation and flux monitor procedures at BGC Samples were irradiated in the OSU TRIGA reactor CLICIT facility in three different irradiation batches: BGC269 in a 14h irradiation and BGC252 and BGC257 in a 20h irradiation. BGC monitors the horizontal flux gradient over 190mm diameter and 2-3 mm thick aluminum discs (S1). All irradiations use FCs. At BGC J-values are calculated by a planar regression over the disc. After rigorous cleaning of the data set (i.e. omission of all data with 40 Ar* yields < 98%, < , Ca/K > 1, followed by omission of data outside 2 standard deviations) weighted mean J values for each hole containing FCs are used for this regression. If standards (and thus also samples) are measured during different periods, J values are calculated for each period and only combined when they are statistically equivalent. Residuals calculated for the standard positions from these regressions are on the order of 0.1%, in practice a slightly higher uncertainty (± 0.2% 1σ uncertainty in J) is arbitrarily used in the age calculations. Irradiations and flux monitor procedures at VU Samples were irradiated in the OSU Triga reactor CLICIT facility in three different irradiation batches all for 7 hours (VU37, VU41 and VU42). The vertical neutron gradient flux is monitored by irradiating samples and standards in ~80mm long vials. The horizontal flux gradient in the 6mm ID vials is ignored. Irradiation VU37 uses Taylor

4 Creek Rhyolite sanidine (TCs) and irradiations VU41 and VU42 use FCs as neutron flux monitor. J-values are calculated using a weighted 2 nd order polynomial regression. Cleaning of the dataset is less rigorous than at BGC, but 40 Ar* yields for FCs and TCs are in general higher than 97%. Radiogenic 40 Ar yields lower than 90%, and Ca/K > 1 are omitted. When MSWD > student s T statistic outliers are omitted until MSWD < student s T (at 95% confidence level). Then weighted mean J values and standard errors for each position in the vial are used in the weighted 2 nd order polynomial regression. Residuals are on the order of 0.1 %, but in practice a slightly higher uncertainty ( % 1σ uncertainty in J) is arbitrarily used in the age calculations. 40 Ar/ analytical procedures at BGC Samples and standards are melted using a focused CO 2 laser to extract trapped argon from the crystals. Gasses are exposed to a -130ºC cold trap to remove H 2 O and to SAES getters to remove reactive compounds. Samples, standards, blanks and air aliquots are measured during a fixed period of 1800 seconds in ~13 scans. Baseline is measured at masses 35.6 and 36.4 at the beginning and end of an analysis. A series of air measurements (n = 4-8) is typically performed at the beginning and end of a series run (after bake out and before sample change). Further single air pipettes are analyzed every ~10 unknowns. Depending on their behavior blanks are measured every one or two unknowns. 40 Ar/ dating procedures at BGC are described in detail elsewhere (S1). 40 Ar/ analytical procedures at VU Samples are melted with a 20W continuous wave argon ion laser. Samples and standards are preheated with a defocused laser beam at ~1.8W to remove contaminating volatiles before melting. Gasses are exposed to SAES getters during 300 seconds, followed by measurement of the argon isotopes on a MAP mass spectrometer. The mass spectrometer is operated with a modified version of the standard MAP software that allows data collection for argon using a Johnston s SEM collector. Measurements are performed in scans in a peak jumping mode from m/e 40 to m/e 35.5 in half mass intervals. Baselines in each scan consist of typically 4-6 integration cycles, 37 Ar and 38 Ar peaks consist of 8-12 integration cycles and 40 Ar, and 36 Ar of integration cycles per scan. The mean value of each integration cycle is stored in a raw data file together with the number of scans and the time elapsed since the inlet of the gas into the mass spectrometer. System blanks are generally measured every 3 or 4 steps. The total system blanks are in the range of moles for mass 40, moles for mass 39, 38 and 36 and moles for mass 37. Mass discrimination ( per atomic mass unit) was monitored by frequent analysis of 40 Ar/ 38 Ar reference gas or 40 Ar/ 36 Ar air pipette aliquots. Further details on 40 Ar/ dating procedures at VU are described in detail elsewhere (S3, S4).

5 Data reduction, age calculation and error propagation 40 Ar/ ages are calculated with decay constants of Steiger and Jäger (S2) and an age of Ma for FCs and Ma for TCs (S1). At the VU the in-house developed ArArCALC program is used for reduction of the raw 40 Ar/ data. Details on procedures and algorithms can be found in Koppers (S5). At BGC the in-house developed argon data acquisition and data reduction program Mass Spec (by A. Deino) is used for data reduction. In both BGC and VU regression lines are calculated, in general by exponential fitting of 40 Ar and and linear fits for 38 Ar, 37 Ar and 36 Ar. Blank correction is done by either a (linear or smoothing) trend through a series of blanks or by bracketing blanks (using linear interpolation). Mass discrimination factors are generally based on linear interpolation. All analytical information is given in the supplementary tables S1, S2. Ages are calculated with 1σ analytical error (i.e. only the error in the 40 Ar * / K ratio of the sample is included) and ± 1σ(J) including the error in the 40 Ar * / K ratio of the standard as well. Outlier detection Approaches in detections of outliers between BGC and VU differ at some points. At BGC data yielding either low radiogenic 40 Ar * (i.e. <97% for unknowns and <98% for FCs), Ca/K ratios > 1 and/or small amounts of (< moles) are immediately discarded. Outliers in the remaining data set are first visually assessed and extreme outliers (i.e. >1 Ma younger/older than expected age) are rejected. Next, a 1.5 nmad criteria is used for rejection of outliers, which appeared to be a very efficient semiquantitative way to recognize outliers. At the VU data yielding either Ca/K ratios > 1 and/or small amounts of (< moles) are also immediately discarded, but criteria for low radiogenic 40 Ar * yields are less rigorous (i.e. data with 40 Ar * < 85% are excluded). Extreme outliers (> 1 Ma from expected age) are then omitted. For the remaining data the 1.5 nmad outlier criteria is applied.

6 Ca/K Age (Ma) Ca/K Age (Ma) Ca/K Mes1: DE F G H I J KL MNO PQ R ST Mes1: I J K L M N O P Q RS T U V W X 6.90 ± 0.02 Ma ± 0.02 Ma Mes4: F G H I J K N 6.75 ± 0.02 Ma Mes4: BC D E F G H I J K L N S 6.75 ± 0.02 Ma Mes12: B C D E F G H I JKL R T ± 0.02 Ma Mes12: N O P Q R S T U V W X ± 0.02 Ma Mes1: Mes4: Mes12: Age (Ma) 6.9 LM N O P Q R S TU 6.88 ± 0.02 Ma 6.9 K L M N O P Q R S TUV 6.4 JK L M N O P Q R S T U V 6.42 ± 0.02 Ma released (%) 6.76 ± 0.02 Ma released (%) released (%) Figure S1. Incremental heating spectra. Incremental heating experiments were performed at BGC on single sanidine crystals of three samples Mes1, Mes4 and Mes12. Analytical data of steps containing less than 2% of the total are listed in the table S2, but are never included in the plateau age. An age plateau is located on basis of MSWD with probability cut-off at 0.95 and the two outermost steps from each side of the plateau must not be significantly different than the weighted mean plateau age at 1.4σ (tolerance limit of 1.4). Radiogenic 40 Ar yields are high (>97% for all steps yielding >2% K ) and Ca/K ratios are constant. The samples do not show complex thermal histories and this implies that total fusions instead op incremental heating experiments are justified (fig. S1). Age spectra are shown with 2 sigma error bars.

7 Table S1: Summary of 40 Ar/ and astronomical ages in the Melilla-Nador Basin. Weighted mean ages with 1σ analytical errors, including analytical uncertainty in J, are given for each irradiation. Data are reported using sanidine from the Fish Canyon Tuff (FCs) as a dating standard with a reference age of Ma and/or sanidine from the Taylor Creek Rhyolite (TCRs) of Ma (S1) and decay constants used by convention since 1977 (S2). Full analytical data and summary calculations for individual irradiation experiments are available in Table S2. Astronomical ages are given in bold. Published 40 Ar/ ages of previous studies on the same tephras (S6-S8) are listed for completeness, but are not included in the intercalibration. Lab ID Irradiation Age (Ma) ± 1σ (J) Prob. n Mes1 >6.84 Ma BGC252A / BGC252A / BGC252A /7 01M0165 VU / BGC252PR-C 6.893* / BGC252PR-C 6.901* / BGC252PR-C 6.880* /22 Mes4 = Ma BGC269A / BGC269B / BGC269C / BGC269D / BGC269E / BGC252A / BGC252A / BGC252A /9 01M0170/191 VU /4 01M0379 VU /10 02M0267 VU / BGC252PR-C 6.753* / BGC252PR-C 6.746* / BGC252PR-C 6.758* /22 V2 (S6) /10 V2 (S6) /10 Me-5 (S8) 6.73* /3 Mes6 = Ma 01M0171 VU /6 Mes8 = Ma BGC257PR-C /15 01M0192 VU /6 02M0268 VU /10 Me-13 (S8) /10 Mes 9 = Ma BGC257PR-C /16 00M0239/245/246 VU /3 01M0167 VU /5 02M0270 VU /10 Me-16 (S8) /10

8 Mes10 = Ma 02M0272 VU /9 Mes11 = Ma BGC257PR-C /30 00M0160 VU /5 Mes12 = Ma BGC252A / BGC252A / BGC252A /8 00M0172/248/168 VU /8 02M0273 VU / BGC252PR-C 6.396* / BGC252PR-C 6.408* / BGC252PR-C 6.419* /22 Mes14 = Ma BGC257PR-C /14 01M0166 VU /8 02M0275 VU /10 Mes15 = Ma 02M0277 VU /8 Mes16 = Ma 02M0279 VU /10 Mes17 = Ma 00M0237/238 VU /3 00M0236 VU /1 02M0280 VU /10 Mes18 = Ma 02M0302 VU /10 Ifo1 = Ma 02M0304 VU /10 Ifo2 = Ma 02M0303 VU /10 Ifo3 = Ma BGC257PR-C /16 01M0110 VU /8 02M0307 VU /10 Ifo4 = Ma BGC257PR-C /17 01M0113/193 VU /7 02M0308 VU /10 Ifo5 = Ma BGC257PR-C /17 01M0112 VU /6 02M0309 VU /10 Ro-3 (S7) /11 Iza = Ma 00M0153/01M0190 VU /11 *weighted mean plateau ages.

9 Table S2: Analytical 40 Ar/ data (pages 9-20). Full analytical data and summary calculations for individual irradiation experiments with 1σ errors. Ages with 1σ analytical errors, including analytical uncertainty in J, are given for each experiment. Data are reported using sanidine from the Fish Canyon Tuff (FCs) as a dating standard with a reference age of Ma and/or sanidine from the Taylor Creek Rhyolite (TCRs) of Ma (S1) and decay constants used by convention since 1977 (S2).

10 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J Single crystal incremental heating - BGC 252PR-C: Mes1 (32456) J = ± A E %40Ar* < B E C E D E E E F E G E H E I E J E K E L E M E N E O E P E Q E R E S E T E PR-C: Mes1 (32456) J = ± A E %40Ar* < B E %40Ar* < C E %40Ar* < D E E E F E G E %40Ar* < H E I E J E K E L E M E N E O E P E Q E R E S E T E U E V E W E X E PR-C: Mes1 (32456) J = ± A E %40Ar* < B E %40Ar* < C E %40Ar* < D E %40Ar* < E E %40Ar* < F E G E %40Ar* < H E I E J E K E L E M E N E O E P E

11 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J Q E R E S E T E U E V E PR-C: Mes4 (32459) J = ± A E %40Ar* < B E C E D E E E F E G E H E I E J E K E L E M E N E O E P E PR-C: Mes4 (32459) J = ± A E %40Ar* < B E C E D E E E F E G E H E I E J E K E L E M E N E O E P E Q E R E S E PR-C: Mes4 (32459) J = ± A E %40Ar* < B E %40Ar* < C E %40Ar* < D E E E F E G E H E I E J E K E L E M E N E O E P E Q E R E S E T E U E V E

12 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J 252PR-C: Mes12 (32459) J = ± A E %40Ar* < B E C E D E E E F E G E H E I E J E K E L E M E N E O E P E Q E R E S E T E PR-C: Mes12 (32459) J = ± A E %40Ar* < B E %40Ar* < C E %40Ar* < D E %40Ar* < E E F E G E H E I E J E K E L E M E N E O E P E Q E R E S E T E U E V E W E X E PR-C: Mes12 (32459) J = ± A E %40Ar* < B E %40Ar* < C E D E E E F E G E H E I E J E K E L E M E N E O E P E Q E R E S E T E

13 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J U E V E Single crystal total fusions - BGC 252A: Mes1 (32474) J = ± Bcomb E Bcomb E Bcomb E Bcomb E Bcomb E E E E A: Mes1 (32480) J = ± Bcomb E Bcomb E Bcomb E E E E E E A: Mes1 (32486) J = ± E Bcomb E Bcomb E E E E E >1.5 nmads from median age 252A: Mes4 (32470) J = ± Bcomb E Bcomb E Bcomb E Bcomb E Bcomb E E E E A: Mes4 (32476) J = ± Bcomb E Bcomb E Bcomb E Bcomb E E E E E A: Mes4 (32482) J = ± Bcomb E Bcomb E E Bcomb E E E E E Bcomb E A: Mes4 (22009) J = ± E E E

14 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J E E E E E E E %40Ar* < E %40Ar* < E %40Ar* < E %40Ar* <97 269B: Mes4 (22029) J = ± B E B E B E B E B E B E B E B E B E B E B E C: Mes4 (22050) J = ± E E E E E E E E E E E E D: Mes4 (22071) J = ± E E E E E E E E E E E E E E E E E E: Mes4 (22083) J = ± E E E E E E E E >1.5 nmads from median age E >1.5 nmads from median age 22083B E age < 6.5 Ma

15 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J 22083B E age < 6.5 Ma 22083B E age < 6.5 Ma 22083B E age < 6.5 Ma 22083B E age < 6.5 Ma 22083B E age < 6.5 Ma 22083B E age < 6.5 Ma 257PR-C: Mes8 (32529) J = ± E E E E E E E E E E E E E E >1.5 nmads from median age E >1.5 nmads from median age 257PR-C: Mes9 (32533) J = ± E E E E E E E E E E E E E >1.5 nmads from median age E age > 7.2 Ma E %40Ar* < E %40Ar* <97 257PR-C: Mes11 (32531) J = ± E E E E E E E E E E E E E E E E E E >1.5 nmads from median age E >1.5 nmads from median age E >1.5 nmads from median age E >1.5 nmads from median age E Ca/K > E %40Ar* < E %40Ar* < E %40Ar* < E %40Ar* <97

16 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J E %40Ar* < E %40Ar* < E %40Ar* < E %40Ar* <97 252A: Mes12 (32472) J = ± Bcomb E Bcomb E Bcomb E Bcomb E Bcomb E E E E A: Mes12 (32478) J = ± Bcomb E Bcomb E Bcomb E Bcomb E E E E E >1.5 nmads from median age 252A: Mes12 (32484) J = ± Bcomb E Bcomb E Bcomb E E E E Bcomb E E >1.5 nmads from median age 257PR-C: Mes14 (32523) J = ± E XX E X E E E E E E E E E E >1.5 nmads from median age E >1.5 nmads from median age E age > 7.2 Ma 257PR-C: Ifo3 (32535) J = ± E E E E E E E E E E E E E E E E >1.5 nmads from median age

17 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (% of ( 10-2 ) ( 10-2 ) ( 10-4 ) (%) excluding including because (W) total) error on J error on J 257PR-C: Ifo4 (32537) J = ± E E E E E E E E E E E E E E %40Ar* < E %40Ar* < E %40Ar* < E %40Ar* <97 257PR-C: Ifo5 (32539) J = ± E E E E E E E E E E E E E E >1.5 nmads from median age E %40Ar* < E %40Ar* < E %40Ar* <97 Single crystal total fusions - VU Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (%) excluding including because (W) error on J error on J VU37: Mes1 (01M0165) J = ± (TCR) 01M0165A E M0165B E M0165D E M0165E E M0165F E M0165H E M0165I E M0165J E VU37: Mes4 (00M0170, 01M0191) J = ± (TCR) 00M0170Brecomb E M0170E E M0170G E M0191A E VU41: Mes4 (B4) J = ± M0379A E M0379B E M0379D E M0379E E M0379F E M0379H E M0379I E M0379J E M0379N E

18 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (%) excluding including because (W) error on J error on J 01M0379K E >1.5 nmads from median age VU42: Mes4 (A3) J = ± M0267A E M0267C E M0267D E M0267E E M0267G E M0267H E M0267I E M0267K E M0267L E M0267M E VU37 Mes6 (01m0171) J = ± (TCR) 01M0171B E M0171D E M0171G E M0171I E M0171K E M0171N E VU37: Mes8 (01M0192) J = ± (TCR) 01M0192A E M0192B E M0192D E M0192E E M0192F E M0192G E >1.5 nmads from median age VU42: Mes8 (A4) J = ± M0268B E M0268C E M0268D E M0268F E M0268G E M0268H E M0268J E M0268K E M0268N E M1267Lrecomb E VU37: Mes9 (B3: 00M0239, 00M0245, 00M0246) J = ± (TCR) 00M0246recomb E M0245recomb E M0239recomb E VU37: Mes9 (01M0167) J = ± (TCR) 01M0167A E M0167C E M0167D E M0167E E M0167G E VU42: Mes9 (A5) J = ± M0270A E M0270B E M0270D E M0270E E M0270F E M0270H E M0270J E M0270L E M0270M E M0270I E >1.5 nmads from median age VU42: Mes10 (A7) J = ± M0272B E M0272C E M0272D E M0272F E M0272G E M0272K E

19 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (%) excluding including because (W) error on J error on J 02M0272H E >1.5 nmads from median age 02M0272J E >1.5 nmads from median age 02M0272L E >1.5 nmads from median age VU37: Mes11 (00m0160) J = ± (TCR) 00M0160A E M0160D E M0160E E M0160F E M0160B E Ca/K >1 VU37: Mes12 (00M0172,00M0248,01M0168) J = ± M0172A E M0172D E M0248D E M0168A E M0168D E M0168E E M0168B E >1.5 nmads from median age 00M0248E E age > 7.2 Ma VU42: Mes12 (A8) J = ± M0273A E M0273B E M0273D E M0273E E M0273F E M0273H E M0273I E M0273J E M0273M E M0273L E >1.5 nmads from median age VU37: Mes14 (01M0166) J = ± (TCR) 01M0166A E M0166B E M0166C E M0166E E M0166F E M0166G E M0166I E M0166J E VU42: Mes14 (A9) J = ± M0275A E M0275C E M0275D E M0275E E M0275G E M0275I E M0275K E M0275L E M0275M E M0275H E >1.5 nmads from median age VU42: Mes15 (A11) J = ± M0277B E M0277D E M0277E E M0277F E M0277H E M0277I E M0277L E M0277M E VU42: Mes16 (A12) J = ± M0279A E M0279C E M0279D E M0279E E M0279G E

20 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (%) excluding including because (W) error on J error on J 02M0279H E M0279I E M0279K E M0279L E M0279M E VU37: Mes17 (00M0237,00M0238) J = ± (TCR) 00M0237B E M0237D E M0238recomb E VU37: Mes17 (00M0236) J = ± (TCR) 00M0236recomb E VU42: Mes17 (A13) J = ± M0280A E M0280B E M0280C E M0280E E M0280F E M0280G E M0280I E M0280J E M0280K E M0280M E VU42: Mes18 (A15) J = ± M0302C E M0302D E M0302E E M0302G E M0302H E M0302I E M0302K E MA302A E MA302C E MA302B E >1.5 nmads from median age VU42: Ifo1 (A17) J = ± M0304A E M0304B E M0304D E M0304E E M0304F E M0304H E M0304I E M0304M E M0304J E >1.5 nmads from median age 02M0304L E >1.5 nmads from median age VU42: Ifo2 (A16) J = ± M0303A E M0303B E M0303C E M0303E E M0303F E M0303I E M0303K E M0303M E M0303G E >1.5 nmads from median age 02M0303J E >1.5 nmads from median age VU37: Ifo3 (01M0110) J = ± (TCR) 01M0110B E M0110D E M0110F E M0110H E M0110K E M0110N E M0110P E M0110Q E

21 Lab ID# Laser 40 Ar/ ± 1σ 38 Ar/ ± 1σ 37 Ar/ ± 1σ 36 Ar/ ± 1σ Ca/K ± 1σ 40 Ar * 40 Ar * / K ± 1σ Age (Ma) ± 1σ ± 1σ Omitted Power (moles) (%) excluding including because (W) error on J error on J VU42: Ifo3 (A19) J = ± M0307B E M0307C E M0307D E M0307F E M0307G E M0307H E M0307J E M0307K E M0307L E M0307N E VU37: Ifo4 (01M0113/01M0193) J = ± (TCR) 01M0193B E M0193C E M0193D E M0193F E M0113C E M0113E E M0193A E >1.5 nmads from median age VU42: Ifo4 (A20) J = ± M0308A E M0308B E M0308D E M0308E E M0308F E M0308H E M0308I E M0308J E M0308L E M0308M E VU37: Ifo5 (01M0112) J = ± (TCR) 01M0112A E M0112D E M0112F E M0112I E M0112K E M0112N E VU42: Ifo5 (A21) J = ± M0309A E M0309C E M0309D E M0309E E M0309G E M0309H E M0309I E M0309K E M0309L E M0309M E VU37: Iza1 (00M0153,01M0190) J = ± (TCR) 01M0190A E M0190B E M0190E E M0190F E M0190G E M0190I E M0153A E Ar * < 50% 00M0153B E >1.5 nmads from median age 00M0153C E >1.5 nmads from median age 00M0153E E >1.5 nmads from median age 01M0190C E >1.5 nmads from median age Isotope ratios are corrected for mass discrimination, blank values and radio-active decay

22 Age and error equations for astronomically calibrated standard T FC 1 = ln λ λt FC 1 astro [( e 1) R 1] = ln[ X ] astro λ [equation 5 in (S1)] TFC 2 TFC 2 T FC σ T = σ σ FC λ T astro FC λ t astro R astro with σ 2 R FC astro TFC λ 1 = ln 2 λ [ X ] R FC astro tastro e λ X λt astro T t R FC astro T FC FC astro = = R FC astro e X λt astro λ astro ( e ) t 1 λ X λ = (5.463 ± 0.107) [decay constant, note that this not the value used by convention, but the value of (S9)] t astro = ± 0.005) 10 6 yr [astronomical age] R FC astro = ± [intercalibration factor between FCs and Melilla tephra, see caption Fig. 2] Above values yield ± Ma for FCs (1 sigma), or ± Ma (2 sigma) Note that the error increases to ± Ma and ± Ma for respectively an uncertainty of ± 10 kyr and ± 20 kyr in the astronomical age (all 1 sigma).

23 Table S3 Published 40 Ar/ data for Matuyama-Bruhnes reversal Dates are reported as in the original publication along with the standard, standard age, dated material and analytical methods. To allow direct comparison all data are recalculated to the astronomically calibrated FCs age of ± Ma (this study). Not all studies report their data relative to FCs. Those data are converted to FCs ages using the following intercalibration factors: R FC TC ± (S1); R FC AC ± (S1); R FC SB [calculated with FC Ma and SB Ma (S10)], R FC FCT-3biotite ± (S11) and R FC Be4-biotite = R FC SB-3 R SB-3 Be4-biotite = = [with Be4-biotite Ma relative to SB-3 biotite (S12)]. Total external uncertainties at the 95% confidence level are reported for recalculated ages. For R FC SB-3 and R FC Be4-biotite uncertainties are not given and are excluded in uncertainties in recalculated ages. The astronomical age for this boundary is 781 kyr (S13). Publication Age (ka) Material Standard Method* Age relative to FCs Ma (in ka) 40 Ar/ dating on lava flows with transitional polarity assuming to be the MB boundary (S14) 783 ± 16 whole rock of 3 flows, Maui, Hawaii FCT3 biotite Ma rel. to SB3-bio wm of 3 isochron ages 794 ± 16 (S15) ± 2.0 whole rock and groundmass of 6 flows, TCs / ACs wm of 6 isochron ages, 20 experiments 780 ± 2 Maui, Hawai (S16) ± 12.4 groundmass of 3 flows, La Palma ACs wm of 3 isochron ages, 14 experiments 803 ± 12 isochron age La Palma flow (S17) (S17) ± 3.8 plagioclase, groundmass and whole rock of 8 flows, on Maui, Tahiti, Chili and La Palma TCs wm of 8 isochron ages 795 ± 4 (S18) ± 3.0 whole rock of 9 flows, Chili TCs / ACs wm of 9 isochron ages, 22 experiments 797 ± 4 40 Ar/ dating on volcanic domes with slightly below the MB boundary (reversed polarity) (S19) 793 ± 18 sanidine of Cerro Santa Rosa II dome, FCs / TCs both rel. to MMhb-1 wm of 4 multigrain laser fusions 812 ± 18 Jemez Mountains, New Mexico (S19) 812 ± 24 sanidine of Cerro San Luis dome, FCs / TCs both rel. to MMhb-1 wm of 9 multigrain laser fusions 831 ± 24 Jemez Mountains, New Mexico (S19) 794 ± 8 sanidine of Serro Seco dome, Jemez FCs / TCs both rel. to MMhb-1 wm of multigrain 4 laser fusions 813 ± 8 Mountains, New Mexico (S20) 787 ± 30 sanidine of Santa Rosa II dome, Valles FCT-3s 27.9 isochron age of 8 single crystal laser fusions 795 ± 30 caldera, New Mexico (S20) 800 ± 6 sanidine of San Luis dome, Valles FCT-3s 27.9 isochron age of 11 single crystal laser fusions 809 ± 6 caldera, New Mexico (S20) 800 ± 14 sanidine of Seco dome, Valles caldera, New Mexico FCT-3s 27.9 isochron age of 14 single crystal laser fusions 809 ± Ar/ dating on volcanic layers bracketing the MB boundary (S19) 789 ± 6 Sanidine of Oldest Toba Tuff, Sumatra FCs / TCs both rel. to MMhb-1 wm of multigrain laser fusion (n=4) 808 ± 6 (reversed polarity) (S21) > 747 ± 12 anorthoclase of airfall ash, Olorgesaillie FCs isochron age of 12 single crystal anorthoclase > 757 ± 12 Fm, Kenya (in normal polarity interval) laser fusions (S22) 791 ± 40 biotite of ash-d ODP site 758 (ash D = 5 cm below MBB) FCT-3 biotite wm of 4 multigrain laser fusions: 800 ± 20 ka for ash-d. Age for MB by linear interpolation of 801 ± 40 (S19) 772 ± 4 sanidine of pumice of upper unit and basal airfall unit Bishop Tuff (S23) ± 3.6 sanidine of air fall pumice and ash flow of Bishop Tuff FCs & TCs both rel. to MMhb * wm = weighted mean; Authors indicate that they probably dated a MB precursor. sedimentation rate wm of 23 multigrain laser fusions: 760 ± 2 ka. Age for MB by adding 12 kyr based on linear interpolation of sedimentation rate TCs Ma rel. to MMhb wm of 69 single crystal laser fusions: ± 1.8 ka. Age for MB by adding 15.3 (± 2.2) kyr based on linear interpolation of sedimentation rate 790 ± ± 4

24 Astronomical tuning of the K-T boundary (figs. S2, S3a-c, S4) Figure S2: Original tuning of (S24) to the eccentricity time series of the Va03_R7 (S25) solution. Dinares-Turell et al. (S24) using the lithological expression of a 2.4 Myr eccentricity minimum (marked by * around 62 Ma) in their cycle bundles as starting point for the tuning. This tuning was extended by matching successive 100-kyr limestone beds to successive 100-kyr eccentricity minima arriving at an K-T boundary age of ~65.8 Ma. The eccentricity time series of the La2004 solution (S26) and corresponding tuning are added for comparison following the same approach with limestone beds as starting point and extending the tuning by matching successive 100-kyr limestone beds to successive 100-kyr eccentricity minima now arriving at an K-T boundary age of ~65.6 Ma. Numbers indicated in the column left of the lithological column denote the successive prominent ~100 kyr eccentricity related cycles of (S24) with limestone beds tuned to eccentricity minima in the middle. Note that the position of the boundaries of the 100-kyr cycles differ slightly from the positions indicated in (S24). Crosses to the right of the lithological column mark mid-points of limestone dominated parts of the 100-kyr cycle that correlate with 100-kyr eccentricity minima. The boundaries between the cycles represent mid-points of the more marly intervals of the 100-kyr cycle that correlate with 100-kyr eccentricity maxima. The magnetostratigraphy ranges from top C29r into C27r. Partially filled black shading denotes two short intervals of normal polarities above C27n, which may represent cryptochrons (S24). The K-T boundary is located at 0 m at the base of this part of the section Polarity E-bundles C29r C29n C28r C28n C27r C27n Lithologic column Va03_R7 La2004 Initial tuning Age (in Ma) * m

25 405-max max max max 100-max p p p 618 0p max max 100-max p max 405-max 405-max Figure S3a. Upper part of the Zumaia section of (S24) below the San Telmo chapel. In the upper photograph, both 100-kyr limestone beds 29 to 42 of (S24) and large scale clusters of precession-related basic cycles that mark successive 405-kyr eccentricity maxima have been indicated. Lower photograph shows the detailed pattern of the precession-related basic cycles (limestone-marl couplet) in the interval that ranges from 100-kyr limestone bed 36 to bed 42. Solid lines mark distinct to prominent marlbeds of the individual precession related cycles, dashed lines mark less distinct to vague marlbeds. Precession cycle numbers (with a -p) correspond to number as in (S24). The phase relations with the 100- and 405-kyr eccentricity cycles have also been indicated (see also caption to Fig. S4).

26 Figure S3b. Middle part of the Zumaia section of (S24). The lower photograph shows the interval ranging from 100-kyr limestone bed 21 to 31 of (S24), with a stratigraphic overlap with the upper part of the section shown in Fig. S3a. Also shown are clusters of well-developed 100-kyr marl beds that mark successive 405-kyr eccentricity maxima. Upper photograph shows the interval ranging from 100-kyr limestone bed 13 to 22 of (S24) and the clustering of intervals with well-developed marly layers that mark successive 405-kyr eccentricity maxima. Red lines mark small faults.

27 59- p T p 11 50p 46 -p p T8 T7 T6 41 -p 34 -p p p 41 -p p p 2 5-p K/T 1 Figure S3c. Lower part of the Zumaia section of (S24) as exposed on the SW side of the Aitzgorri headland. The upper photograph shows the interval ranging from 100-kyr limestone bed 10 to 13 of (S24). Also shown is the cluster of well-developed marly layers between 100-kyr limestone beds 10 and 12 (or 13) that corresponds to a single 405-kyr eccentricity maximum. Patterns are less obvious due to the intercalation of several turbidites (T6-9). Lower photograph shows the interval ranging from the K-T boundary up to the 100-kyr limestone bed 12 of (S24) and the clustering of intervals with well-developed marly layers that corresponds to successive 405-kyr maxima. On both photographs, numbers of some characteristic precession-related cycles of (S24) are indicated (with a p).