Deep-sea record of impact apparently unrelatedto mass extinction in the Late TriassicTetsuji Onouea,1, Honami Satoa, Tomoki Nakamurab, Takaaki Noguchic, Yoshihiro Hidakad, Naoki Shiraid,Mitsuru Ebiharad, Takahito Osawae, Yuichi Hatsukawae, Yosuke Tohe, Mitsuo Koizumie, Hideo Haradae,Michael J. Orchardf, and Munetomo Nedachig

aDepartment of Earth and Environmental Sciences, Kagoshima University, Kagoshima 890-0065, Japan; bDepartment of Earth and Planetary Material Sciences,Tohoku University, Miyagi 980-8578, Japan; cDepartment of Science, Ibaraki University, Mito 310-8512, Japan; dDepartment of Chemistry, Tokyo MetropolitanUniversity, Tokyo 192-0397, Japan; eQuantum Beam Science Directorate, Japan Atomic Energy Agency (JAEA), Ibaraki 319-1195, Japan; fGeological Surveyof Canada, Vancouver, BC, Canada V6B 5J3; and gDivision of Instrumental Analysis, Frontier Science Research Center, Kagoshima University, Kagoshima890-0065, Japan

Edited by Dennis V. Kent, Rutgers University/Lamont-Doherty Earth Observatory, Palisades, NY, and approved October 3, 2012 (received for reviewJune 4, 2012)

The 34-million-year (My) interval of the Late Triassic is marked bythe formation of several large impact structures on Earth. LateTriassic impact events have been considered a factor in bioticextinction events in the Late Triassic (e.g., end-Triassic extinctionevent), but this scenario remains controversial because of a lack ofstratigraphic records of ejecta deposits. Here, we report evidencefor an impact event (platinum group elements anomaly withnickel-rich magnetite and microspherules) from the middle Norian(Upper Triassic) deep-sea sediment in Japan. This includes anom-alously high abundances of iridium, up to 41.5 parts per billion(ppb), in the ejecta deposit, which suggests that the iridium-enriched ejecta layers of the Late Triassic may be found on a globalscale. The ejecta deposit is constrained by microfossils that suggestcorrelation with the 215.5-Mya, 100-km-wide Manicouagan impactcrater in Canada. Our analysis of radiolarians shows no evidence ofa mass extinction event across the impact event horizon, and nocontemporaneous faunal turnover is seen in other marine plank-tons. However, such an event has been reported among marinefaunas and terrestrial tetrapods and floras in North America. We,therefore, suggest that the Manicouagan impact triggered theextinction of terrestrial and marine organisms near the impact sitebut not within the pelagic marine realm.

chert | Panthalassa | Mino Terrane | pelagic sediments

The discoveries of an iridium-enriched clay layer at the Cre-taceous/Paleogene (K/Pg) boundary and the 65-million-year

(My)-old Chicxulub impact structure in Mexico have revealedthe global environmental damage such impacts can cause andtheir association with mass extinction events (1, 2). Analysis ofthe K/Pg event has led to the assumption that other large ex-traterrestrial impacts also contributed to mass extinction eventsover the past 250 million years (3, 4). The Late Triassic, 236–201million years ago (Mya), is marked by several large impact cra-ters (5): the 100-km Manicouagan and the 40-km Saint Martinstructures in Canada; the 25-km Rochechouart structure inFrance; the 10-km Paasselkä structure in Finland; and the 9-kmRed Wing structure in the United States (Fig. 1). Previousresearchers suggested that these Late Triassic impact craterscould have resulted in the end-Triassic extinction event (6, 7) orearlier extinctions at the Norian/Rhaetian or Carnian/Norianboundaries (8, 9). To understand the causal link between theLate Triassic extinctions and the impact events, it is necessary toprecisely date the age of crater formation and locate the geologicrecord of the impact ejecta layer within a well age-constrainedstratigraphic sequence. Although the ages of the Late Triassicimpact events have been constrained by radioisotopic dating (SIAppendix, Table S1), deposits containing their ejecta are knownonly from a nonmarine sequence in southwestern Britain (10–12).Here, we report that Upper Triassic deep-sea sediment (bedded

chert) at Sakahogi in the Inuyama area, Central Japan, contains

microspherules, nickel (Ni)-rich magnetite, and a platinum groupelement (PGE) anomaly, all of which suggest an extraterrestrialimpact event (2, 13). A previous paleomagnetic study of theSakahogi locality indicated that the bedded chert accumulated atlow latitudes (3.9° ± 3.3° N or S) within a deep Paleo-Pacific(Panthalassa) basin (14, 15). We examine the high-resolutionbiostratigraphy of microfossils across the impact event horizonand discuss the relationship between the impact and extinctionevents in the Late Triassic deep-sea sediments.

Impact Ejecta LayerEvidence of a Late Triassic impact event was discovered asa PGE anomaly from a claystone layer in an Upper Triassicbedded chert succession of the Sakahogi locality, Inuyama area,Central Japan (SI Appendix, Fig. S1). The claystone layer, whichwe call the “Sakahogi ejecta layer,” contains microspherules, Ni-richmagnetite, and a high abundance of PGEs, ranges in thicknessfrom 4 to 5 cm, and extends laterally for at least 90 m at Sakahogi.We distinguish lower and upper sedimentary sublayers withinthe claystone (SI Appendix, Fig. S3). The lower sublayer (∼8-mm-thick) contains 10–15% (by rock volume) microspherules ina matrix of clay minerals (mainly illite), cryptocrystalline quartz,and hematite. Our preliminary investigation reveals that thissublayer contains many small euhedral to subhedral crystals ofoxidized Ni-rich magnetite (16). Geochemical mapping usinga scanning X-ray analytical microscope also reveals a high con-centration of nickel (SI Appendix, Fig. S4). The upper sublayer ofthe claystone is composed of undisturbed sediments of clay (illite)and cryptocrystalline quartz containing radiolarians, siliceoussponge spicules, and conodonts.The geochemical signals of the extraterrestrial impact are

recorded in the lower sublayer of the claystone. Anomalies iniridium and other PGEs in the lower sublayer were identifiedbased on analyses of nine claystone and two chert samples byinstrumental neutron activation analysis (INAA) and inductivelycoupled plasma–mass spectrometry (ICP–MS). The anomalouslyhigh abundances of iridium, which is limited to the lower sub-layer (Fig. 2), is defined by concentrations of up to 41.5 parts perbillion (ppb), much greater than the background level of ∼0.2

Author contributions: T. Onoue, H.S., T. Nakamura, T. Noguchi, M.E., T. Osawa, andY. Hatsukawa designed research; T. Onoue, H.S., T. Nakamura, T. Noguchi, Y. Hidaka,N.S., M.E., T. Osawa, Y. Hatsukawa, Y.T., M.K., H.H., M.J.O., and M.N. performed research;T. Nakamura contributed new reagents/analytic tools; T. Onoue, H.S., T. Nakamura,T. Noguchi, Y. Hidaka, N.S., M.E., T. Osawa, Y. Hatsukawa, Y.T., M.K., H.H., M.J.O.,and M.N. analyzed data; and T. Onoue and H.S. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: onoue@sci.kagoshima-u.ac.jp.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1209486109/-/DCSupplemental.

19134–19139 | PNAS | November 20, 2012 | vol. 109 | no. 47 www.pnas.org/cgi/doi/10.1073/pnas.1209486109

ppb (SI Appendix, Table S2). The magnitude of this concentra-tion is comparable to that measured at K/Pg sites (17), suggestingthat the iridium-enriched ejecta layers of the Late Triassic maybe found in sites worldwide (18). Anomalies in other PGEs arealso seen in the lower sublayer of the claystone (Fig. 2). Fig. 3shows the average elemental abundances of PGEs (relative to CI

carbonaceous chondrite) at four stratigraphic levels in the studysection. In the lower sublayer, the concentrations of all of theseelements are well above the background values in the underlyingand overlying levels, and are enriched by up to three orders ofmagnitude compared with average terrestrial crustal abundances(19). The PGE anomaly was confirmed by an additional analysis of

200

210

220

230

Age

(Myr)

Stage/

Substage

Hettangian

Carnian

Lower

Middle

Upper

Rhaetian

Impact crater

A BEjecta

depositFZ

4B

5A

5B

6A

6B

8C

8D

0A

8B8A7

0B

30

LandShelfDeep water

Impact siteEjecta deposit

Japan

P (10 km)228.7 ± 3 Myr

SW Britain

Fig. 1. Late Triassic paleogeography and age of impact craters and ejecta deposits. (A) Approximate locations of five impact craters and ejecta sites plottedon a Late Triassic paleogeographic map modified from the original figure in ref. 5. The hatched area indicates the inferred depositional area of the beddedchert from the Mino Terrane in the low-latitudinal zone of the Panthalassa Ocean (14). See B for abbreviations for the craters and ejecta sites. (B) Radio-isotopic and stratigraphic ages of the five impact craters based on the timescale proposed by ref. 27. Triassic radiolarian fossil zones (FZ) are from ref. 45. Notethat the age of the ejecta deposit is constrained by conodonts and radiolarians (base of the 6B fossil zone) and is, therefore, given without an error range. Theblue shading in the fossil zone column indicates the total age range of the section shown in Fig. 5.

0

10

20cm

Chert

Xip

hosp

haer

afis

tula

ta

Syr

ingo

caps

aba

tode

s

Cap

nodo

cesa

risa

Cap

nodo

cecr

ysta

llina

Japo

noca

mpe

nova

Cor

umre

gium

Stic

hoca

psa

cf. n

ana

Cap

nodo

ceru

esti

Po u

lpus

piab

yx

Cap

nuch

osph

aera

cf.d

ewev

eri

Tria

latu

sro

bust

us

10 20 30 40

Mid

dle

No

ria

n

Up

pe

rT

ria

ssi c

6A

Ep i

gon d

olel

labi

dent

ata

E. p

oste

r a/ E

.se r

r ula

ta

AgeDistribution of PGEs (ppb)Species ranges of radiolarian fossils

0

Ejecta layer

CZRZ

Cap

nodo

cecf

.ana

pete

s

Cap

nodo

ceex

tent

a

53.0

52.4

52.352.252.1

51.251.0

50.2

Os (ICP-MS)

Ir (ICP-MS)

Ru (ICP-MS)

Pt (ICP-MS)

6B

SampleNo.(NH-)

Ir (INAA)

Fig. 2. PGE abundances and biostratigraphy of radiolarians from the middle Norian section (Sakahogi) in Japan. Solid squares beside the lithologic sectionindicate the occurrence of conodonts and radiolarians in cherts and claystones. The radiolarian ranges are used to constrain the age of the ejecta deposit.Radiolarian (RZ) and conodont (CZ) zones are from refs. 45 and 22, respectively. See SI Appendix, Tables S2 and S3 for PGE data.

Onoue et al. PNAS | November 20, 2012 | vol. 109 | no. 47 | 19135

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

the second sample (NH-52R2) from the lower sublayer in theSakahogi locality (SI Appendix, Table S3). Because PGEs are highlydepleted in the Earth’s crust relative to solar abundances (19, 20),the PGE anomaly reported here may have resulted from the ac-cretion of a significant amount of siderophile elements from a largeprojectile. However, additional sampling for PGEs is needed tofurther test the origin of the projectile and to estimate its size.The discovery of microspherules and Ni-rich magnetite asso-

ciated with the PGEs anomalies in the lower sublayer also sug-gests an impact event (2, 13, 21), although they are affected bysecondary processes, including low-grade regional metamor-phism of the study area [conditions below those of the prehnite–pumpellyite facies (15)], which altered their chemical composition.Microspherules were only found in the lower sublayer and havenever been reported in the Triassic bedded chert succession of theInuyama area. The microspherules in the lower sublayer are darkgreen to dark gray and range in size from 200 to 300 μm (Fig. 4A).Synchrotron X-ray diffraction analysis indicates that the micro-spherules are pseudomorphs of clinochlore-rich chlorite (SI Ap-pendix, Fig. S5). These microspherules preserve a pseudomorphicinternal texture that comprises dendrites and spherulites radiatingfrom the surface into the microspherules (Fig. 4B). Some micro-spherules contain a high proportion of small, euhedral to sub-hedral crystals of oxidized Ni-rich magnetite (Fig. 4C). The Ni-richmagnetite grains are 5–20 μm in size and commonly have skeletalor octahedral morphologies (Fig. 4D). Electron microprobe anal-ysis shows large variations in the concentrations of Al, Fe, Cr, Ni,and Zn among the magnetite grains, along with minor amounts ofMg, Ti, and Mn (SI Appendix, Table S4). These magnetite grainsare distinguished from typical igneous magnetite by high contentsof Ni and Fe3+ and relatively low concentrations of Ti.

Age and BiostratigraphyThe depositional age of the claystone is constrained by radiolariansand conodonts (SI Appendix, Table S6 and Fig. S6). A detaileddiscussion of their biostratigraphy in the Sakahogi locality is givenin SI Appendix, SI Text. The claystone layer lies between pelagicchert layers of the radiolarian zones TR6A and TR6B and is cor-related with the upper middle Norian. A biostratigraphic analysisof conodonts also reveals that the ejecta deposit is embedded in theupper middle Norian (base of the Epigondolella bidentata zone ofref. 22). This age indicates that the impact event occurred wellbefore the Rhaetian interval (Epigondolella mosheri and Mis-ikella posthernsteini conodont zones), and, therefore, it has no

relevance to the extinction events at the end-Triassic and Nor-ian/Rhaetian boundaries (7, 23, 24).Our biostratigraphic analysis suggests that there was no mass

extinction of radiolarians across the impact event horizon. Fig. 5shows biostratigraphic ranges of 147 species in 62 genera ofradiolarians from 331 horizones; the collection intensity wasuniform without any sampling gaps (SI Appendix, Fig. S2). Asignificant faunal turnover is observed ∼1 m above the impactejecta horizon in the upper middle Norian. Given that the av-erage sedimentation rate of the middle Norian chert succession,estimated from the measured thickness and the time interval ofits deposition (SI Appendix, Fig. S2), is 1.0 mmper thousand years,this turnover occurred ∼1 My after the impact event. Only onespecies became extinct at the ejecta horizon and the extinctionrate of radiolarians (extinct species divided by total species at thesame level) is estimated to be about 5% at the horizon. High-resolution radiolarian biostratigraphy across the impact ejectalayer also indicates that, of the 13 radiolarian species identifiedbelow this horizon, only one species,Trialatus robustus, disappearsat the horizon (Fig. 2). The extinction of the genus Trialatusappears to have occurred synchronously across several regions inthe middle Norian (SI Appendix, SI Text), meaning its last oc-currence may be a good time indicator for the ejecta deposit.

DiscussionA biostratigraphic analysis of radiolarians and conodonts revealsthat the ejecta deposit is embedded in the upper middle Norian.Recent magnetostratigraphic studies linked to conodont bio-stratigraphy (25–27) suggest that the stratigraphic position of theSakahogi ejecta layer can be correlated with paleomagneticchron PM9 to PM10 at the Pizzo Mondello section in Sicily (26)and with SB-9 at the Silická Brezová section in Slovakia (25).Although no radiometric age data are available, the magneto-stratigraphic ages from these Norian sections have been assignedan age range of 216–212 Mya (SI Appendix, Fig. S7) based on thecorrelation with the astronomically tuned geomagnetic polaritytime scale (Newark-APTS of refs. 28–30). These ages wouldseem to consistent with the dating of an impact ejecta layer insouthwestern Britain that has yielded a diagenetic age (from

Os Ir Ru Pt Rh Pd10-5

10-4

10-3

10-2

10-1

1

K/Pg boundary

Upper continental crust

Norian ejecta deposit

Norian deep-sea deposit

NH-52.1

NH-52.4

NH-52.3

NH-52.2 (ejecta deposit)

CI-

norm

aliz

edP

GE

abun

danc

e s

Fig. 3. CI chondrite–normalized PGE patterns of middle Norian samples atSakahogi. Stratigraphic levels of samples are shown in Fig. 2. The concen-trations of CI chondrites are from ref. 20. PGE patterns of the upper conti-nental crust (19) andK/Pg boundary sediments (17) are shown for comparison.See SI Appendix, Table S3 for data.

Fig. 4. Microspherules and Ni-rich magnetite from a Norian ejecta depositin Japan. (A) Photomicrographs of microspherules from the ejecta deposit.Plane-polarized light. (Scale bar: 500 μm.) (B) Microspherule showing theinward-radiating growth form of clinochlore pseudomorphs. (Scale bar: 100μm.) (C) Scanning electron micrographs (backscattered mode) of a polishedsection of microspherule from the ejecta deposit. Ni-rich magnetite (brightgrains) commonly occur as skeletal grains. (Scale bar: 100 μm.) (D) Close-upview of the microspherule in C illustrating skeletal crystal morphologies ofNi-rich magnetite in the microspherule. (Scale bar: 20 μm.)

19136 | www.pnas.org/cgi/doi/10.1073/pnas.1209486109 Onoue et al.

authigenic K-feldspar) of 214 ± 2.5 My (11), although therecalculated age of 216.7 My for the British ejecta layer using themethod of Renne and coworkers (31, 32) is slightly outsidethe range of 216–212 Mya. However, the 40Ar/39Ar age of theBritish ejecta layer has yet to be confirmed by a second analysis,and the preferred age of ∼214 My is based on a plateau age thatincludes only slightly more than 50% of the gas released (11).Accurate age dating of the ejecta layer, coupled with an im-proved understanding of the correlation of the radioisotopic ageswith the magnetobiostratigraphy of the Norian stage (25–27),would be required to precisely compare the ages between theejecta layers of Japan and southwest Britain. The mineral as-semblage of the spherules from the Sakahogi locality differ fromthose of hollow illitic and glauconitic spherules reported fromsouthwestern Britain (10, 11). However, in many cases, impactspherules in ejecta deposits are devitrified and altered to claysresulting in loss of original chemical composition (33). There-fore, the difference in spherule composition between Japan andsouthwest Britain spherules does not rule out the possibility ofthe same impact origin for these spherules.Based on recent radioisotopic ages for Late Triassic impact

structures, there are two candidate impacts that may have pro-duced the ejecta found in the Sakahogi deposit: the 100-kmManicouagan structure in northeastern Canada [dated at 215.5Mya (34)] and the 9-km Red Wing structure in North Dakota,United States (dated at 200 ± 25 Mya; SI Appendix, Table S1).We suggest that the Manicouagan impact structure is the sourceof the Sakahogi ejecta deposit, because (i) the age of the Man-icouagan crater formation is consistent with the age of the ejectadeposit and (ii) the Manicouagan impact was large enough toproduce a global distribution of ballistic ejecta within ∼30°paleolatitude of the crater.* If the projectile and target rocks of

the Manicouagan impact (8) could have produced the PGE-enriched ejecta layer in Japan, the pelagic Paleo-Pacific paleo-position of the ejecta deposit suggests that impact ejecta layerswill be found at other middle Norian sites worldwide. Sites atwhich the ejecta horizon might be found include pelagic lime-stone sequences of the Pizzo Mondello in Sicily (26) and BlackBear Ridge in British Columbia (22) and the nonmarine sequenceof the Newark Supergroup at the eastern margin of NorthAmerica (28–30). The application of an event-stratigraphic ap-proach (13) at such sections is required to confirm the distributionof the Manicouagan ejecta. It is also important to determinewhether the absolute age of these middle Norian sections corre-lates with the 215.5-Mya Manicouagan impact.Because the middle Norian is now known to contain evidence

of an impact event, an analysis of the extinction patterns ofmarine and terrestrial biotas at a more regional scale might bethe next step in investigating the nature of extinction and bi-ological turnover events at ejecta horizons. Marine microfossilshave been thought to be one of the broad taxonomic groups mosteffected by the Chicxulub impact event at the K/Pg boundary (2,35). Our analysis of radiolarians does not show a mass extinctionevent across the impact ejecta layer (Fig. 5). Along with radio-larians, dinoflagellates and calcareous nannoplanktons are thedominant marine plankton in the Late Triassic ocean (36, 37).On a generic level, a survey of Upper Triassic dinoflagellate cystand calcareous nannoplankton assemblages shows neither a massextinction nor a marked decline across the impact event horizon(SI Appendix, Fig. S8). The record of their floral change in theLate Triassic has been established at the species level in severalregions (36, 37), and the record shows that there was no ex-tinction event around the impact event horizon. Therefore, wesuggest that many groups of marine planktons survived theManicouagan impact event. Other marine fossil records to datesupport the absence of globally synchronous mass extinctions inthe middle Norian (38, 39). However, late middle Norian marineextinctions have been reported at the genus and/or species level

Age RZ

8C8B

8A7

6A5B

Rha

etia

nC

arni

anU

pper

Nor

ian

Low

erN

or. M

iddl

eN

or.

Upp

erT

riass

ic

226.

62 1

6.4

2 11.

820

9.8

6B

Species ranges of radiolarians in Inuyama area

Ejecta horizon

0

10

20

30

(m)

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

35

Carnian

Norian

NorianRhaetian

201

(Mya

)

Het

t.

Jura

.

8D40

Rhaetian

Hettangian0A0B

Fig. 5. Stratigraphic ranges of Late Triassic radiolarian species in Inuyama area projected onto the composite section. Species numbers are shown on the xaxes. For an explanation of radiolarian taxon ranges, see SI Appendix, Table S5. Radiolarian zone (RZ) is from ref. 45.

*Wrobel KE, Schultz PH, Thirty-Fourth Annual Lunar and Planetary Science Conference,March 17–21, 2003, Houston, TX, abstr 1190.

Onoue et al. PNAS | November 20, 2012 | vol. 109 | no. 47 | 19137

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

in ammonoids, bivalves, and conodonts from western NorthAmerica (40). It is possible that the Manicouagan impact mayhave triggered the extinctions of these marine faunas, at least inwestern North America, but this will require further biostrati-graphic study at middle Norian sites globally.Terrestrial records of Late Triassic vertebrates and plants have

been well studied in the Chinle Formation of the Petrified ForestNational Park in southwestern United States. Recent magneto-stratigraphic and radioisotopic studies of the Chinle Formationsuggest that a faunal turnover of terrestrial tetrapods, includingthe disappearance of herbivorous dicynodonts, occurred in themiddle Norian (41–43). This turnover may coincide with a floralturnover in North America deduced from palynology studies(41, 42, 44). The depositional site of the Chinle Formation isrelatively close to the Manicouagan crater (within ∼3,500 km),raising the possibility that the effects of the impact [e.g., acid rainand wildfire on a regional scale (35)] triggered a catastrophicdisruption of the terrestrial ecosystems in North America (42).However, it is uncertain whether the igneous and metamorphictarget rocks of Manicouagan impact (8) could have producedcatastrophic environmental effects (35), and a lack of age dataprevents confirmation of the concurrence of the Manicouaganimpact and floral and faunal turnovers in North America (41,43). Detailed studies of terrestrial sections in the middle Norianare needed to determine whether the biotic turnover of terres-trial biotas at this time was caused by the Manicouagan impact.

Materials and MethodsFor whole-rock analyses of PGEs, eleven samples were powdered in an agatemortar. Veins and strongly recrystallized zones were avoided to minimize theeffects of diagenetic or metamorphic overprinting. PGE abundances weredetermined using inductively coupled plasma–mass spectrometry combinedwith a nickel sulfide fire assay (SI Appendix, SI Materials and Methods).Iridium was measured using a neutron activation technique. Powder sampleswere enclosed in small pure quartz vials and irradiated for 48 h in the JRR-3reactor at the Japan Atomic Energy Agency (Ibaraki, Japan). Iridium wasquantified with a gamma–gamma spectrometer detecting the 468- to 316-keV γ-ray coincidence resulting from the decay of 192Ir.

We handpicked microspherules from the crushed rock powders (<500-μmgrain size) of the ejecta deposit under a binocular microscope. To investigatethe mineral composition of individual microspherules, synchrotron X-raydiffraction analysis was performed using a Gandolfi camera with synchro-tron X rays of 2.165-Å wavelength at the High Energy Accelerator ResearchOrganization (Tsukuba, Japan). After this analysis, microspherules weremounted in epoxy, sectioned, and polished to reveal their internal textureusing a field-emission scanning electron microscope. Chemical compositionsof Ni-rich magnetite were determined by electron microprobe with anelectron acceleration voltage of 15 keV and a beam current of 10 nA. Detailsof our sample preparation and analytical methods can be found in SI Appendix,SI Materials and Methods.

ACKNOWLEDGMENTS. We thank C. Yasuda for assistance in the field; theHigh Energy Accelerator Research Organization (KEK) for synchrotron X-raydiffraction analysis; and the Japan Atomic Energy Agency (JAEA) forinstrumental neutron activation analysis. This work was supported by theJapan Society for the Promotion of Science.

1. Alvarez LW, Alvarez W, Asaro F, Michel HV (1980) Extraterrestrial cause for the cre-taceous-tertiary extinction. Science 208(4448):1095–1108.

2. Schulte P, et al. (2010) The Chicxulub asteroid impact and mass extinction at theCretaceous-Paleogene boundary. Science 327(5970):1214–1218.

3. Alvarez W, Muller RA (1984) Evidence from crater ages for periodic impacts on theEarth. Nature 308(5961):718–720.

4. Raup DM, Sepkoski JJ, Jr. (1984) Periodicity of extinctions in the geologic past. ProcNatl Acad Sci USA 81(3):801–805.

5. Spray JG, Kelley SP, Rowley DB (1998) Evidence for a late Triassic multiple impactevent on Earth. Nature 392(6672):171–173.

6. Bice DM, Newton CR, McCauley S, Reiners PW, McRoberts CA (1992) Shocked quartzat the triassic-jurassic boundary in Italy. Science 255(5043):443–446.

7. Olsen PE, et al. (2002) Ascent of dinosaurs linked to an iridium anomaly at the Triassic-Jurassic boundary. Science 296(5571):1305–1307.

8. Hodych JP, Dunning GR (1992) Did the Manicouagan impact trigger end-of-Triassicmass extinction? Geology 20(1):51–54.

9. Tanner LH, Lucas SG, Chapman MG (2004) Assessing the record and causes of LateTriassic extinctions. Earth Sci Rev 65(1–2):103–139.

10. Kirkham A (2003) Glauconitic spherules from the Triassic of the Bristol area, SWEngland: Probable microtektite pseudomorphs. Proc Geol Assoc 114:11–21.

11. Walkden G, Parker J, Kelley S (2002) A late Triassic impact ejecta layer in southwesternBritain. Science 298(5601):2185–2188.

12. Thackrey S, et al. (2009) The use of heavy mineral correlation for determining thesource of impact ejecta: A Manicouagan distal ejecta case study. Earth Planet Sci Lett285(1–2):163–172.

13. Koeberl C (2001) The sedimentary record of impact events. Accretion of Extraterres-trial Matter Throughout Earth’s History, eds Peucker-Ehrenbrink B, Schmitz B (KluwerAcademics, New York), pp 333–378.

14. Ando A, Kodama K, Kojima S (2001) Low-latitude and Southern Hemisphere origin ofAnisian (Triassic) bedded chert in the Inuyama area, Mino terrane, central Japan.J Geophys Res 106(B2):1973–1986.

15. Matsuda T, Isozaki Y (1991) Well-documented travel history of Mesozoic pelagic chertin Japan: From remote ocean to subduction zone. Tectonics 10(2):475–499.

16. Sato H, Onoue T (2010) Discovery of Ni-rich spinels in Upper Triassic chert of the MinoTerrane, central Japan. J Geol Soc Jpn 116(10):575–578.

17. Ganapathy R (1980) A major meteorite impact on the Earth 65 million years ago:Evidence from the cretaceous-tertiary boundary clay. Science 209(4459):921–923.

18. Claeys P, Kiessling W, Alvarez W (2002) Distribution of Chicxulub ejecta at the Cre-taceous-Tertiary boundary. Spec Pap Geol Soc Am 356:55–69.

19. Peucker-Ehrenbrink B, Jahn BM (2001) Rhenium-osmium isotope systematics andplatinum group element concentrations: Loess and the upper continental crust. Ge-ochem Geophys Geosyst 2:2001GC000172.

20. Anders E, Grevesse N (1989) Abundances of the elements: Meteoritic and solar. Ge-ochim Cosmochim Acta 53(1):197–214.

21. Glass BP, Simonson BM (2012) Distal impact ejecta layers: Spherules and more. Ele-ments 8(1):43–48.

22. Orchard MJ (1991) Upper Triassic conodont biochronology and new index speciesfrom the Canadian Cordillera. Geol Surv Can Bull 417:299–335.

23. Olsen PE, Shubin NH, Anders MH (1987) New early Jurassic tetrapod assemblagesconstrain Triassic-Jurassic tetrapod extinction event. Science 237(4818):1025–1029.

24. Sephton MA, et al. (2002) Carbon and nitrogen isotope disturbances and an end-

Norian (Late Triassic) extinction event. Geology 30(12):1119–1122.25. Channell JET, et al. (2003) Carnian-Norian biomagnetostratigraphy at Silicka Brezova

(Slovakia): Correlation to other Tethyan sections and to the Newark Basin. Palae-

ogeogr Palaeoclimatol Palaeoecol 191(2):65–109.26. Muttoni G, et al. (2004) Tethyan magnetostratigraphy from Pizzo Mondello (Sicily)

and correlation to the Late Triassic Newark astrochronological polarity time scale.

Geol Soc Am Bull 116(9–10):1043–1058.27. Hüsing SK, Deenen MHL, Koopmans JG, Krijgsman W (2011) Magnetostratigraphic

dating of the proposed Rhaetian GSSP at Steinbergkogel (Upper Triassic, Austria):

Implications for the Late Triassic time scale. Earth Planet Sci Lett 302(1–2):203–216.28. Kent DV, Olsen PE (1999) Astronomically tuned geomagnetic polarity time scale for

the Late Triassic. J Geophys Res 104(B6):12831–12841.29. Olsen PE, Kent DV (1999) Long-period Milankovitch cycles from the Late Triassic and

Early Jurassic of eastern North America and their implications for the calibration of

the Early Mesozoic time-scale and the long-term behaviour of the planets. Phil Trans

Royal Soc London Ser A 357(1757):1761–1786.30. Olsen PE, Kent DV, Whiteside H (2010) Implications of the Newark Supergroup-based

astrochronology and geomagnetic polarity time scale (Newark-APTS) for the tempo

and mode of the early diversification of the Dinosauria. Earth Environ Sci Trans R Soc

101:201–229.31. Renne PR, Mundil R, Balco G, Min K, Ludwig KR (2010) Joint determination of 40K

decay constants and 40Ar*/40K for the Fish Canyon sanidine standard, and improved

accuracy for 40Ar/39Ar geochronology. Geochim Cosmochim Acta 74(18):5349–5367.32. Renne PR, Balco G, Ludwig KR, Mundil R, Min K (2011) Response to the comment by

W.H. Schwarz et al. on “Joint determination of 40K decay constants and 40Ar*/40K for

the Fish Canyon sanidine standard, and improved accuracy for 40Ar/39Ar geo-

chronology” by P.R. Renne et al. (2010). Geochim Cosmochim Acta 75(17):5097–5100.33. French BM, Koeberl C (2010) The convincing identification of terrestrial meteorite

impact structures: What works, what doesn’t, and why. Earth Sci Rev 98(1–2):123–170.34. Ramezani J, Bowring SA, Pringle MS, Winslow FD III, Rasbury ET (2005) The Man-

icouagan impact melt rock: A proposed standard for the intercalibration of U-Pb and40Ar/39Ar isotopic system. Geochim Cosmochim Acta 60(10):A321.

35. Kring DA (2007) The Chicxulub impact event and its environmental consequences at

the Cretaceous-Tertiary boundary. Palaeogeogr Palaeoclimatol Palaeoecol 255(1–2):

4–21.36. Bown PR (1992) Late Triassic-Early Jurassic calcareous nannofossils of the Queen

Charlotte Islands, British Columbia. J Micropalaeont 11(2):177–188.37. Riding JB, Mantle DJ, Backhouse J (2010) A review of the chronostratigraphical ages

of Middle Triassic to Late Jurassic dinoflagellate cyst biozones of the North West Shelf

of Australia. Rev Palaeobot Palynol 162(4):543–575.38. Benton MJ (1986) More than one event in the late Triassic mass extinction. Nature 321

(6073):857–861.39. Simms MJ, Ruffell AH (1990) Climatic and biotic change in the late Triassic. J Geol Soc

London 147(2):321–327.40. Orchard MJ, et al. (2001) The Upper Triassic of Black Bear Ridge, Williston Lake,

northeast British Columbia: An integrated biochronology. Geol Surv Can Curr Res A6:

1–21.

19138 | www.pnas.org/cgi/doi/10.1073/pnas.1209486109 Onoue et al.

41. Irmis RB, Mundil R, Martz JW, Parker WG (2011) High-resolution U-Pb ages from theUpper Triassic Chinle Formation (New Mexico, USA) support a diachronous rise ofdinosaurs. Earth Planet Sci Lett 309(3–4):258–267.

42. Parker WG, Martz JW (2011) The Late Triassic (Norian) Adamanian-Revueltian tetra-pod faunal transition in the Chinle Formation of Petrified Forest National Park, Ari-zona. Earth Environ Sci Trans R Soc Edinburgh 101:231–260.

43. Ramezani J, et al. (2011) High-precision U-Pb zircon geochronology of the Late Tri-assic Chinle Formation, Petrified Forest National Park (Arizona, USA): Temporal

constraints on the early evolution of dinosaurs. Geol Soc Am Bull 123(11–12):2142–2159.

44. Litwin RJ, Traverse A, Ash SR (1991) Preliminary palynological zonation of the Chinleformation, southwestern U.S.A., and its correlation to the Newark supergroup(eastern U.S.A.). Rev Palaeobot Palynol 68(3–4):269–287.

45. Sugiyama K (1997) Triassic and Lower Jurassic radiolarian biostratigraphy in the sili-ceous claystone and bedded chert units of the southeastern Mino Terrane, CentralJapan. Bull Mizunami Fossil Mus 24:79–193.

Onoue et al. PNAS | November 20, 2012 | vol. 109 | no. 47 | 19139

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S