A rain of ordinary chondritic meteorites in the early ... · orite of the project. From this bed...

A rain of ordinary chondritic meteorites in theearly Ordovician

Birger Schmitz a;*, Mario Tassinari b, Bernhard Peucker-Ehrenbrink c

a Marine Geology, Earth Sciences Centre, P.O. Box 460, SE-405 30 Go«teborg, Swedenb Va«ner Museum, P.O. Box 724, SE-531 17 Lidko«ping, Sweden

c Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA

Received 23 July 2001; received in revised form 12 October 2001; accepted 15 October 2001

Abstract

Forty fossil meteorites with a total original mass of V7.7 kg have been recovered in the first systematic search forfossil meteorites, pursued in an active quarry in Lower Ordovician (480 Ma) marine limestone in southern Sweden. Themeteorites represent at least 12 different falls over a seafloor area of V6000 m2 during 91.75 Myr, making the quarryone of the most meteorite dense areas known in the world. Geochemical analyses of relict chromite grains indicate thatall or most of the meteorites are ordinary chondrites and probably L chondrites. Mechanisms for meteorite deliveryfrom the asteroid belt to Earth were the same 480 Ma as today, however, the flux was one to two orders of magnitudehigher, most likely reflecting the disruption of the L chondrite parent body at about that time. This is a major event inlate solar-system history, which may also have led to an enhanced flux of asteroids to Earth during V30 Myr. ß 2001Elsevier Science B.V. All rights reserved.

Keywords: meteorites; ordinary chondrites; chromite; asteroids; Os-188/Os-187

1. Introduction

Modern in£ux rates and compositional varia-tions of meteorites on Earth are well-constrained.Sky-watch camera networks and meteorite searchprograms in wind-eroded deserts indicate that onaverage between 36 and 116 meteorites s 10 gaccumulate annually per 106 km2 of the Earthsurface [1,2]. Meteorite fall statistics indicate

that 80% of the meteorites are ordinary chon-drites (H, L or LL), 7.9% achondrites, 4.8% ironmeteorites, and 4.6% carbonaceous or anomalouschondrites [3]. Enstatite chondrites and stonyirons make up only 1.6% and 1.1% of the falls,respectively. Among recent ordinary chondritesthe H and L groups make up 45% each, withthe remainder being LL chondrites. For the dis-tant geological past we know hardly anythingabout the meteorite in£ux to Earth. Prior to thestart of this project in 1992 only half a dozenfossil meteorites were known (see [4]). This is avery small number considering that during the lastcentury billions of tons of coal and limestone wereindustrially quarried and vast areas of sedimenta-

0012-821X / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 1 ) 0 0 5 5 9 - 3

* Corresponding author. Tel. : +46-31-7734902;Fax: +46-31-7734903.

E-mail addresses: [email protected] (B. Schmitz),[email protected] (B. Peucker-Ehrenbrink).

EPSL 6057 13-12-01

Earth and Planetary Science Letters 194 (2001) 1^15

www.elsevier.com/locate/epsl

ry rocks have been inspected by ¢eld geologists.The rarity of fossil meteorites led authors in themid-20th century to argue that meteorites did notfall before the late Quaternary [4]. Meteorite ex-perts, like H.H. Nininger, counterargued that me-teorites would not survive in recognizable form inolder rocks and that no systematic attempt torecover fossil meteorites had been performed [5].

In a preliminary account of our systematic fos-sil meteorite search in the Thorsberg quarry(58³35PN, 13³26PE) at Kinnekulle, southern Swe-den, we reported 17 meteorite ¢nds [6]. Here wepresent the ¢rst detailed documentation and eval-uation of our ¢ndings. With 40 meteorites found,from the end of 1992 to the end of 2000, our database is now su¤ciently large to make ¢rm quan-titative estimates of the meteorite abundance onthe Ordovician sea£oor. We also present a sum-mary of the chemical analyses of relict chromitegrains from 26 of the fossil meteorites, aimed atconstraining the types of meteorites that fell 480Ma. We discuss sedimentological and chemicalapproaches to the problem of pairing the fossilmeteorites, critical for estimating £uxes. More de-tailed reports of the chromite analyses, the search

process, and the meteorites found will be givenelsewhere.

The active part of the Thorsberg quarry spans asection of 3.2 m of Lower Ordovician marinelimestone (Figs. 1 and 2). Twelve prominentbeds, 11^62 cm thick, can be discerned; eachhas a name traditionally used by the quarry work-ers. The beds occur in a horizontal position andhave not been a¡ected by tectonism. The lowerV80 cm and the upper V1 m of the sectionconsist of red limestone, separated by V1.4 mgray limestone. The quarried Orthoceratite Lime-stone was deposited during the Arenigian to Llan-virnian in an epicontinental sea that covered sev-eral 100 000 km2 of the Baltoscandian Shield [7,8].This condensed limestone formed at an averagerate of one to a few mm/kyr, however, depositionwas variable, changing from long periods of non-deposition and hardground formation to morerapid pulses of sedimentation [7,8]. Biostrati-graphically, the quarried section belongs to theAmorphognatus variabilis^Microzarkodina £abel-lum conodont subzone [9]. Based on an averagesedimentation rate of V2 mm/kyr for the Areni-gian limestone at Kinnekulle the section is esti-

Fig. 1. The Thorsberg quarry in June 1999. Note water ¢lled basin in front from which the Arkeologen bed has been recovered.

EPSL 6057 13-12-01

B. Schmitz et al. / Earth and Planetary Science Letters 194 (2001) 1^152

mated to represent a time span of 91.75 Myr [9].According to traditional/regional stratigraphic

nomenclature the section in the Thorsberg quarrybelongs to the Lower Ordovician, however, wenote that the International Commission on Ordo-vician Stratigraphy has suggested to revise thede¢nition of the Lower/Middle Ordovicianboundary [10], implying that the section wouldinstead be in the Middle Ordovician series.

The di¡erent beds and sublayers in the Thors-berg quarry are of di¡erent industrial quality. A

major part of the section consists of high-qualitylimestone used for the production of sawed platessold as £oor plates, window sills, stair cases etc.(Fig. 2). Some intervals of lower quality can occa-sionally yield sawed plates, but are primarily usedas raw garden plates or crushed for production ofcement or agricultural lime. Some intervals areonly used for production of crushed rock. Blocksof limestone (ca. 1.25U2U0.5 m) are recoveredand transported to the nearby sawing factory.The quarrying is performed in two steps: ¢rstthe beds above the basal Arkeologen bed are re-moved and thereafter the Arkeologen bed, that isunder the ground water table, is quarried in smallbasins from which the water has to be pumped(Fig. 1). From the end of 1992 to the end of 2000an area of V6000 m2 of the beds overlying theArkeologen bed and V2700 m2 of the Arkeolo-gen bed were quarried. The workers and ownersof the quarry and sawing factory have beentrained by us to recognize meteorites, which areusually found during sawing, but are occasionallyalready discovered on exposed surfaces in thequarry. We visit the quarry and factory regularlyin order to follow and document the industrialprocess and to collect the meteorites recovered.

The fossil meteorites can readily be identi¢edby macroscopical examination (Figs. 3^5), despitebeing completely pseudomorphosed primarily bycalcite, barite and phyllosilicates [9,11,12]. Theonly relict mineral phases are chromite and chro-mian spinel [11,12]. For all 14 meteorites fromlayers above the Arkeologen bed and for 16 of26 meteorites from the Arkeologen bed visualidenti¢cation was con¢rmed by bulk-meteoriteplatinum group element (PGE) [6,9] or osmiumisotope analyses and/or element analyses of relictchromite grains ([6], this study).

2. Chemical and isotopic methods

The element analyses of chromite grains wereperformed using a LINK Oxford energy disper-sive spectrometer with a Ge detector, mounted ina Zeiss DSM 940 scanning electron microscope.Chromite grains were polished to a £at surfaceusing 1 Wm diamond slurry. A cobalt standard,

Fig. 2. Quarried beds and meteorite distribution in theThorsberg quarry. Column at the right shows the intra-beddistribution of the meteorites. The exact positions of the me-teorites from the Tredje Karten bed and most of the meteor-ites from the Arkeologen bed are not known (see main text).Only half of the section is used for production of plates, therest being crushed and not searched for meteorites.

EPSL 6057 13-12-01

B. Schmitz et al. / Earth and Planetary Science Letters 194 (2001) 1^15 3

linked to simple oxide and metal standards, wasused to monitor drift of the instrument. Acceler-ating voltage of 25 kV, sample current about 1 nAand counting live-time of 100 s were used. Preci-sion (reproducibility) of analyses was typicallybetter than 1^4%. Analytical accuracy was con-trolled by repeated analyses of the USNM117075 (Smithsonian) chromite reference standard[13].

For Os isotopic analysis a 10^50 mg samplefrom respective meteorite was spiked with 190Os,homogenized and mixed with a mixture of Ni, S,and borax. The mixture was fused at 1020³C for90 min, cooled and the NiS bead separated anddissolved in hot 6.2 N HCl. The solution waspassed through a 0.45 Wm ¢lter to capture theinsoluble, PGE-containing particles [14]. The ¢lterpaper was dissolved in 1 ml concentrated HNO3

at about 100³C overnight. Osmium was extractedfrom this solution using the sparging technique[15] and analyzed by magnetic sector inductivelycoupled plasma-mass spectrometry (ICP-MS)(Finnigan Element). Rhenium concentrationswere measured on separate 25^95 mg powdered

sample splits. After spiking with 185Re and disso-lution with mineral acids, Re was separated andpuri¢ed on a column of 1 ml of AG1X8 (200^400mesh) in 0.5 N HNO3 and eluted with 8 N HNO3.Subsequently Re was measured by ICP-MS usingconventional liquid uptake. Total Re blanks ofV0.5 pg are negligible compared to the amountof sample Re (300 pg^18 ng).

3. Distribution and sizes of fossil meteorites

The 40 meteorites vary in cross-section from0.7U1 cm to 15U20 cm (Fig. 2). Cross-sectionsof the 26 meteorites from the basal, red Arkeol-ogen bed range from 0.7U1 cm to 7.5U9 cm. Asingle large meteorite, 6U9U2 cm, has beenfound in the overlying Golvsten bed. In the grayinterval only the Botten and Flora beds haveyielded meteorites. From the Botten bed four me-teorites have been recovered, one with a a size ofV15U20U6.5 cm, representing the largest mete-orite of the project. From this bed derive alsoanother large, 5.5U9U3 cm, and two medium-

Fig. 3. The Oë sterplana Sex 001 meteorite together with a nautiloid shell in a limestone plate sawed parallel to sea£oor surface.Note relict chondrule structures and partly peeled-o¡ fusion crust.

EPSL 6057 13-12-01


sized meteorites, 4U5 cm and 3.3U6.5 cm. TheFlora bed has yielded one small meteorite, 2U2.5cm. In the upper red interval six meteorites, from1.4U3 cm to 6.5U8 cm, have been found in theSextummen bed and two small meteorites, 1.3U2cm and 3U3.4 cm, were recovered from theTredje Karten bed. As the volume of limestonequarried per year and the rate of ¢nds have re-mained relatively constant over the past 8 yr weinfer that meteorites were distributed relativelyevenly throughout the quarry.

Most of the meteorites originate from intervalsused for production of sawed plates, which mayprimarily re£ect that the sawing procedure re-quires much more visual inspection of the rockthan other production methods. Comparing thenumber of meteorites recovered from each bedwith the total amount of rock area of respectivebed through which the saw in the factory haspenetrated gives some indications about real var-iations in the abundance of meteorites throughthe section (Table 1). The quarry owners record

Fig. 4. The Oë sterplana Bot 003 meteorite in a limestone plate sawed perpendicularly to sea£oor surface. This is the largest mete-orite found with an original mass of 3.4 kg. Note relict chondrule structures in meteorite and two prominent hardground surfacesbelow the meteorite.

Table 1Fossil meteorite abundance and areas of sawed plates from the Thorsberg quarry

Bed Sawed plates 1993^2000 Meteorites 1992^2000

(m2) (rel.%) (n) (rel.%)

Tredje Karten 0 0.0 2 5.0Sextummen 6 260 7.7 6 15.0Ro«dkarten 5 596 6.9 0 0Flora 7 842 9.6 1 2.5Likhall 16 800 20.6 0 0Gra®karten 13 314 16.3 0 0Botten 11 280 13.9 4 10.0Golvsten 3 750 4.6 1 2.5Arkeologen 16 616 20.4 26 65.0Total 81 458 100.0 40 100.0

EPSL 6057 13-12-01


sawed-area data to determine the lifetime of thesaw blades. Therefore, these data can only giveinformation about the density of meteorites per

rock surface exposed by sawing. The data showthat the Arkeologen bed is signi¢cantly richer inmeteorites than the other sawed beds. It has

Fig. 5. Large plate (70U110 cm) of limestone sawed parallel to sea£oor surface. Several nautiloid shells accumulated on a hard-ground surface in the vicinity of the Oë sterplana Ark 023 meteorite (3.5U4.5 cm). The iron in the red limestone around the mete-orite has been reduced, therefore a halo of lighter gray limestone has formed.

EPSL 6057 13-12-01


yielded 65% of the meteorites but represents only20.4% of the total sawed rock area. No meteoriteshave been recovered from the Likhall and Gra®-karten beds in the middle of the gray interval,although they represent as much as 36.9% of thesawed rock area. These di¡erences re£ect varia-tions in meteorite in£ux and/or sedimentationrate, however, the small di¡erences between otherbeds may not be statistically signi¢cant. The factthat some of the meteorites, such as the two in theTredje Karten bed, have been found in situ in thequarry adds uncertainty to the comparisons inTable 2.

Meteorites typically lie on hardgrounds, wherethey accumulated together with other large ob-jects such as abundant nautiloid shells (Figs. 3^5). This indicates that meteorites were concen-trated due to sediment winnowing by bottom cur-rents, analogous to the process that leads to me-teorite-enriched surfaces in wind-eroded deserts.Fragile nautiloid shells show no preferred orien-tation and are commonly well-preserved, rulingout that meteorites were transported to this areaby strong bottom currents (Fig. 5). Sometimes themeteorites are associated with marl layers (Fig. 4).We have established a detailed understanding ofthe succession of hardgrounds and marl layerswithin the di¡erent beds. In many cases it hasbeen possible to identify the speci¢c horizon atwhich individual meteorites deposited. The Botten

bed, for example, shows several prominent hard-ground surfaces, and there is ¢rm evidence thatthe four meteorites in that bed occur at three dis-tinctively di¡erent surfaces (Figs. 2 and 4). Thered meteorite-yielding beds (Arkeologen, Sextum-men and Tredje Karten) show lateral intra-bedvariations and complex layering which makes itmore di¤cult to assign meteorites to speci¢c hori-zons. In the Arkeologen bed all except one of the26 meteorites appear to derive from a 25 cm thickinterval extensively used for sawing (Fig. 2).While some of these meteorites de¢nitely comefrom di¡erent levels within this interval, detailedpositions for the majority of the meteorites arenot known. The meteorites in the Sextummenbed come from at least four and possibly ¢vedi¡erent horizons. The exact positions of theTredje Karten meteorites are not known.

We estimate that individual meteorites from theArkeologen bed had original masses between V1and V760 g. The total meteorite mass from thisbed is V2.2 kg (Fig. 6). The meteorites from theoverlying beds weighed between V5 and V500 g,except the largest meteorite recovered, thatweighed V3.4 kg (Fig. 4). The total meteoritemass from beds above the Arkeologen bed isV5.5 kg. Masses have been calculated fromcross-section data, assuming an ellipsoidal shape,and using a meteorite density of 3.6 g/cm3. Theheight of the meteorites (when not known) is as-sumed to be 75% of the shortest cross-section inthe bedding plane.

Fig. 6. Mass distribution of fossil meteorites. A schematic il-lustration of the shape of the size distribution curve for 578paired meteorites from the Nullarbor region is also shown[3].

Table 2Os, Re and 187Os/188Os in fossil meteoritesa

Meteorite Os Re 187Os/188Os(ng/g) (ng/g)

Oë sterplana Ark 003 385 þ 0.7 7.2 0.1269 þ 0.0003Oë sterplana Ark 008 733 þ 2 46.9 0.1288 þ 0.0003Oë sterplana Ark 019 169 þ 0.4 0.1262 þ 0.0004Oë sterplana Ark 027 72 þ 0.7 0.1275 þ 0.0010Oë sterplana Gol 001 651 þ 3 4.2 0.1255 þ 0.0005Oë sterplana Bot 002 852 þ 8 737 0.1534 þ 0.0006Oë sterplana Bot 003 1031 þ 5 0.1751 þ 0.0003Oë sterplana Flo 001 94 þ 1 0.1264 þ 0.0009Oë sterplana Sex 001 74 þ 0.2 0.1290 þ 0.0008Oë sterplana Sex 003 1280 þ 5 0.1263 þ 0.0002Oë sterplana Tre 002 163 þ 0.3 0.1270 þ 0.0004a 2c uncertainties are listed for 187Os/188Os and Os based oncounting statistics. In-run precision of the Re analyses was2^3%.

EPSL 6057 13-12-01


Most meteorites found today in search pro-grams in Antarctica and wind-eroded hot desertsweigh less than 100 g and have a peak in the massdistribution at about 10^20 g (Fig. 6) [3,16,17].The median weight for all the fossil meteorites isV40 g, and there are signi¢cantly more large(v100 g) than small (10^20 g) meteorites com-pared to meteorite collections from recent hot de-serts and Antarctica (Fig. 6). This indicates thatour search is biased towards larger meteorites.The 14 meteorites from the beds that overlie theArkeologen bed have a median mass of V100 g,re£ecting that most of these beds are not searchedfor meteorites with as much scrutiny as the Ar-keologen bed.

4. Os, 187Os/188Os, and Re results

Ordinary chondrites that fell on Earth recentlytypically have Os concentrations in the range 420^1050 ng/g and 187Os/188Os ratios of 0.1265^0.1305[18]. The 12 fossil meteorites analyzed in thisstudy show Os concentrations in the range 72^1280 ng/g and 187Os/188Os ratios of 0.1255^0.1751 (Table 2). Seven of the meteorites haveOs concentrations as high as recent meteorites,385^1280 ng/g, whereas ¢ve have lower concen-trations, 74^169 ng/g, most likely re£ecting lossof PGEs during diagenesis. Ten of the fossil me-teorites, all from red limestone or in the red^graytransition interval, have 187Os/188Os ratios in therange 0.1255^0.1290, very similar to recent mete-orites. Two meteorites from the Botten bed, in thegray limestone interval, have 187Os/188Os ratios of0.1534 and 0.1751, signi¢cantly higher than forrecent chondrites. This enrichment in 187Os canbe explained by diagenetic accumulation of Refrom pore solutions in the reducing environmentof the gray sediments, followed by decay of 187Reto 187Os. This is supported by a substantiallyhigher Re concentration, 737 ng/g, in one ana-lyzed meteorite from the Botten bed, comparedto three meteorites from the red or reddish grayintervals, showing Re concentrations in the range4.2^46.9 ng/g (Table 2). This is consistent withprevious ¢ndings that the gray limestone isstrongly enriched in Re and 187Os compared to

the surrounding red limestone [6]. `Initial' 187Os/188Os ratios for the fossil meteorites at 480 Ma aredi¤cult to calculate because of uncertaintiesabout the timing and amount of diagenetic lossor gain of Re. Our estimates of `initial' 187Os/188Os based on realistic assumptions about Remobility support that the Os is of chondritic ori-gin.

5. Meteoritic chromite and its composition

Studies of the two fossil meteorites Brun£o andOë sterplana, found in Ordovician limestone inSweden before our project started, have shownthat the chemical composition of relict chromitecan be used to classify the meteorites [11,12]. Wehave performed elemental analyses of chromitegrains from 15 of the largest meteorites from theArkeologen bed and 11 of the meteorites fromoverlying beds. In addition, chromite grainsfrom a variety of recent ordinary chondriteswere analyzed using identical methods in orderto obtain optimally comparable data. For eachmeteorite a total of 10^40 analyses were per-formed on typically 10^20 `coarse chromite'grains (0.05^0.3 mm; see [19]). The recent chon-drites (n = 33) analyzed constitute a representativesample of groups (H, L and LL) and petrologictypes (3^6). In most ordinary chondrites chromiteis a common trace mineral (0.05^0.5 wt%), where-as it is rare or absent in carbonaceous and ensta-tite chondrites as well as iron meteorites [19^23].

In all fossil meteorites inspected, chromitegrains are common and have chemical composi-tions indicative of ordinary chondrites (Fig. 7;Table 3). Chromites from carbonaceous chon-drites, achondrites, pallasites, mesosiderites andiron meteorites have compositions markedly dif-ferent than chromites from ordinary chondrites[24^29]. The fossil chromite grains show, with afew exceptions, only little variation in MgO, TiO2,Al2O3 and V2O5 within a meteorite and betweendi¡erent meteorites. Chromite FeO, MnO andZnO concentrations are also quite uniform inmost fossil meteorites, but in some meteoritesconcentrations are anomalous and variable, re-£ecting most likely diagenetic e¡ects. There is no

EPSL 6057 13-12-01


Fig. 7. Chemical composition (averages) of chromite grains from 26 fossil and 33 recent meteorites. The following recent ordinarychondrites were analyzed. H3: Brown¢eld 1937; H4: Bath, Dimmitt, Kesen; H5: Abajo, Forest City, Hessle, Jilin, Pultusk; H6:Archie, Benoni, Estacado, Kernouve, Ozona, Mills; L3: Julesburg; L4: Barratta, Saratov, Waltman; L5: Ausson, Elenovka, Et-ter, Tsarev; L6: Baroti, Holbrook, Mbale (L5/6), Modoc 1905, New Concord, Walters; LL5: Altameen; LL6: Beeler, Bison,Dhurmsala.

EPSL 6057 13-12-01


signi¢cant di¡erence in composition betweenchromites from the Arkeologen meteorites andthose from the overlying layers.

Previous work has shown that there are di¡er-ences in chromite composition between recent H,L and LL chondrites of petrologic types 5 and 6,however, the compositional ranges between thegroups overlap [24,25]. Chromites from ordinarychondrites of lower petrologic types show largercompositional variability. Our analyses of chro-mite in recent chondrites con¢rm trends shownin earlier work, such as generally higher FeOand TiO2 and lower MgO and Al2O3 in L andLL than H type 5/6 chondrites [24,25]. Fig. 7shows that the majority of the fossil meteoritesplot in the low MgO and Al2O3 and high TiO2

¢elds indicative of recent L and LL chondrites.For FeO, ZnO and MnO the situation is morecomplicated. Some fossil meteorites have chromitegrains with (on average) much lower FeO (as lowas 18.7 wt%) and much higher ZnO (up to 8.5wt%) and MnO (up to 1.8 wt%) concentrationsthan in recent ordinary chondrites (Table 3; Fig.7). The negative correlation between FeO andZnO (as well as MnO) in some relict chromites(Fig. 7f) and the similarity of the FeO+MnO+ZnO content in recent and fossil chromite (Fig.7c) indicates that FeO can be replaced by ZnOand MnO during diagenesis. Some relict chro-mites have higher Cr2O3 and lower FeO+MnO+ZnO concentrations than recent chromite,which may also be a diagenetic e¡ect (Fig. 7e).The chromites in all fossil meteorites have thesame narrow range in V2O5 concentrations as inrecent ordinary chondrites (Fig. 7d).

Schmitz et al. [6] suggested that the abundanceof meteorites in the Thorsberg quarry may berelated to a high meteorite in£ux following thedisruption of the L chondrite parent body V500Ma [30^32]. The chromite data presented heresupport, but do not unequivocally prove, thatall or a large majority of the fossil meteoritesare indeed L chondrites. Due to some overlap inchromite composition between di¡erent ordinarychondrite groups and the possibility that diage-netic processes have a¡ected relict chromites, amixed origin from three ordinary chondritic par-ent bodies as observed today cannot be ruled out.T

able

3A

vera

geel

emen

tco

ncen

trat

ion

(wt%

)in

chro

mit

efr

omfo

ssil

and

rece

ntm

eteo

rite

sa

Met

s/an

alb

Cr 2

O3

Al 2

O3

MgO

TiO

2V

2O

5F

eOM

nOZ

nOT

otal

Rec

ent

H5/

6gr

oup

11/2

6156

.7þ

0.3

6.45

þ0.

143.

03þ

0.14

2.17

þ0.

160.

88þ

0.02

29.1

2þ

0.42

1.02

þ0.

040.

33þ

0.05

99.7

0R

ecen

tL

5/6

grou

p10

/209

56.0

þ0.

75.

86þ

0.36

2.73

þ0.

642.

73þ

0.42

0.91

þ0.

0230

.16

þ1.

180.

84þ

0.10

0.31

þ0.

0499

.54

All

foss

ilm

eteo

rite

s26

/594

57.6

þ1.

35.

53þ

0.29

2.57

þ0.

832.

73þ

0.40

0.89

þ0.

0426

.94

þ3.

891.

01þ

0.33

1.86

þ2.

4399

.13

Ark

eolo

gen

bed

15/3

3257

.2þ

1.2

5.55

þ0.

262.

35þ

0.37

2.79

þ0.

420.

88þ

0.03

26.9

7þ

4.13

1.05

þ0.

382.

33þ

3.01

99.1

2O

ther

beds

11/2

6258

.2þ

1.3

5.51

þ0.

342.

87þ

1.16

2.64

þ0.

380.

91þ

0.05

26.8

9þ

3.73

0.96

þ0.

251.

21þ

1.15

99.1

9U

SNM

1170

75st

anda

rd(n

=5)

61.6

þ0.

79.

68þ

0.10

15.2

þ0.

370.

11þ

0.03

n.a.

13.0

4þ

0.55

n.a.

n.a.

99.6

3

n.a.

=no

tan

alyz

ed.

aD

ata

give

nar

eav

erag

es(w

ith

S.D

.)of

aver

ages

ofco

mpo

siti

onof

typi

cally

10^2

0ch

rom

ite

grai

nsin

each

met

eori

te.

bN

umbe

rof

met

eori

tes

from

whi

chch

rom

ite

was

anal

yzed

/tot

alnu

mbe

rof

anal

yses

ofch

rom

ite

grai

ns.

EPSL 6057 13-12-01


Moreover, the existence of currently unproductiveor no longer existing parent bodies resembling H,L and LL parent bodies in composition cannot beexcluded either. Oxygen isotopic analyses of relictchromite are in progress and may help better con-straining the origin of the fossil meteorites.

6. Early Ordovician meteorite in£ux rates

The major di¤culty in estimating the £ux ofmeteorites in the early Ordovician from our datais the pairing problem. A meteorite may break upduring £ight, producing several fragments on theground. In reconstructions of present in£ux rates,e.g. in the Nullarbor desert, textural features, cos-mic-ray exposure ages, terrestrial residence ages,and spatial distribution of meteorites can be usedto determine which fragments belong to the samefall [3], but the Ordovician meteorites are too al-tered to be paired in this way. Fossil meteoriteson di¡erent hardgrounds, however, must repre-sent di¡erent falls, because post-depositional ver-tical migration across hardgrounds is very un-likely. Upward migration due to shake-sorting[33], for instance, would result in unrealisticallylong exposure times on the sea£oor. In dry terres-trial environments meteorites disintegrate alreadywithin 20^30 kyr [2,3], however, the rate at whichmeteorites decay on the sea£oor is not known.The well-preserved, often angular, shapes ofmost of our meteorites indicate that they havenot spent any substantial time drifting across thesea£oor (Figs. 3^5). The 40 meteorites were recov-ered from at least 12 di¡erent surfaces and thusrepresent at least 12 falls (Fig. 2). The main un-certainty in the number of independent surfaces isthe di¤culty associated with laterally correlatinghardgrounds in the meteorite-rich Arkeologenbed.

Our data yield a reliable minimum estimate ofthe £ux of meteorites to Earth at 480 Ma. Weknow that at least 40 meteorites with a mass of7.7 kg representing at least 12 falls with individualmasses ranging from 14 to 3400 g fell within asea£oor area of V6000 m2 during 91.75 Myr.It could be argued that uncertainties in the aver-age sedimentation rate (V2 mm/kyr) pose a prob-

lem for the £ux estimates. However, biologicalevolutionary changes over the quarried interval,e.g. for conodonts and trilobites [34,35], are smalland rule out a signi¢cantly longer period of dep-osition than the 1.75 Myr postulated here. Thebest estimate for the present meteorite £ux toEarth is 80 þ 40 meteorites s 10 g per 106 km2

of the Earth's surface per year [1,2], which corre-sponds to only 0.8 þ 0.4 meteorites per 6000 m2

per 1.75 Myr. Our search, however, is biased to-wards ¢nding mainly the larger meteorites. There-fore a comparison with recent £ux in the corre-sponding size range is more meaningful. Based onestimates of the recent £ux 0.28 meteorite s 100 gincluding 0.12 meteorite s 700 g should havebeen recovered from the quarried area [1,36]. As-suming that fossil meteorites on the same hard-ground represent fragments of a single fall impliesthat at least seven meteorites s 100 g includingthree s 700 g fell in the quarried area. These dataindicate that the meteorite £ux was at least 25times higher in the Ordovician than today.

The pairing problem can be circumvented bycomparing total meteorite mass rather than num-ber of falls. Bland [2] calculated that today 5.7^14.3 kg of meteoritic material in the mass range101^103 g accumulates per 106 km2 per year,which is equivalent to 0.06^0.15 kg per 6000 m2

per 1.75 Myr. Correcting our mass estimate forthe fact that the Arkeologen bed has only beensearched for meteorites over 2700 m2 and assum-ing that the remaining 3300 m2 of this bed is asrich in meteorites as the searched part gives atotal mass of 10.4 kg (7.7 kg+2.7 kg). A compar-ison with recent mass estimates implies a factor of69^173 higher meteorite £ux during the Ordovi-cian.

We consider these minimum estimates for thefollowing reasons. The 40 meteorites recoveredrepresent only a fraction of the meteorites thatfell within the area quarried. Only half of thelimestone section is being used for sawed platesand hence carefully examined for meteorites (Fig.2). Even for the beds suitable for production ofsawed slabs not all of the material removed hasbeen inspected for meteorites. Some material maybe sold to other sawing factories or, if there is nodemand for material from a particular bed, the

EPSL 6057 13-12-01


quarried blocks are crushed or stored. Even in thefraction of rock handled in an optimal way formeteorite recovery, a large fraction of the smallmeteorites are not recovered as indicated by thesize distribution data. Many stone products fromthe quarry are thicker than the typical thicknessof meteorites and it is a matter of chance whetherthe saw cuts through a meteorite so that a su¤-ciently large area is exposed for recognition. It isalso possible that only a fraction of meteoritesthat fell on the sea£oor remain preserved in arecognizable state. Sea£oor destruction of meteor-ites most likely occurred at hardgrounds exposedto corrosive bottom waters for long times.

The extremely high meteorite density in the Ar-keologen bed (one meteorite per 100 m2) remainspuzzling. Such densities have only been reportedfor a few, extremely fragment-rich strewn ¢elds,such as that of the Pultusk meteorite shower [37].The available data do not allow us to di¡erentiatewhether the Arkeologen meteorites re£ect astrewn ¢eld or several distinct falls during a peri-od of particularly high meteorite in£ux. Studies ofcoeval strata worldwide may give an answer tothis.

7. Delivery of meteorites from the asteroid belt

Systematic searches for fossil meteorites fromdi¡erent geological periods can be used to testproposed models for meteorite delivery to Earth.For example, the predominance of ordinary chon-dritic material (80% of recent falls) among presentmeteorites is in marked contrast to its rarity(91%) in micrometeorite and stratospheric extra-terrestrial dust samples, believed to be more rep-resentative of the dust-producing asteroid popula-tion [38^41]. This discrepancy led to thehypothesis that asteroids of ordinary chondriticcomposition were never abundant in the main as-teroid belt and that, historically, carbonaceousmaterial has dominated the entire main belt [41].Our ¢ndings do not support this hypothesis, andinstead indicate that the £ux of ordinary chon-dritic meteorites was even more important 480Ma than today.

It has been known for a long time that among

the ordinary chondrites that reach Earth today alarge fraction of L chondrites show 39Ar^40Ar gasretention ages of V500 Ma [30^32]. This indi-cates that one of the major asteroid break-upevents in late solar-system history occurred atabout that time. There is a signi¢cant uncertaintyin the proposed break-up age, and 87Rb^87Sranalyses and a laser-heating Ar dating yield agesof 450^465 Ma [42^44]. Considering the large ageuncertainties, the meteorite-rich sediments in theThorsberg quarry may well have formed just afterthe disruption event. Zappala et al. [45] haveshown that the £ux of asteroids to Earth mayincrease substantially after major asteroid break-up events. More than 80% of the impacts onEarth related to such an event should take placeduring a 2^30 Myr period. The cratering rate dur-ing this period could be enhanced by one to twoorders of magnitude. Depending on the type oforbital resonances involved, there may be a timelag of up to ca. 100 Myr between the break-upevent and the resulting asteroid shower, but inmost situations the lag is less than one or a fewmillion years [45]. Major break-up events can sim-ilarly be expected to lead to a surge of meteoritesto Earth on time scales of 1^10 Myr [45,46].

The correspondence between the age of the Lchondrite parent-body disruption and the manyfossil meteorites (probably L chondrites) in theThorsberg quarry prompted us to scrutinize theterrestrial cratering record for further evidencefor such a disruption event. In Fig. 8 all datedimpact craters (9630 Ma) on Earth are plottedas a histogram with 30 Myr bins [47^49]. The450^480 Ma bin shows a factor of three to fourmore craters than any other 30 Myr bin from 180to 630 Ma. We note also that of 17 known craterson the Baltoscandian Shield four have been un-equivocally dated by chitinozoan stratigraphy tothe 450^480 Ma period (the Tva«ren, Ka«rdla,Lockne and Granby craters) [48,49]. The peak incrater ages at 480^450 Ma could be a re£ection ofthe L chondrite disruption event, however, werealize that the terrestrial crater record is veryincomplete, representing only a fraction of a per-cent of all the craters that have formed on Earth[47]. The assigned ages of many craters also havelarge uncertainties. The observed peak in the

EPSL 6057 13-12-01


abundance of craters with ages 450^480 Ma maythus be purely coincidental. Even if this observa-tion is an artifact, the existing data do not pre-clude that a major asteroid disruption occurred. Abreak-up event in the asteroid belt may lead to alarger relative increase in the £ux of meteorite-sized than asteroid-sized objects to Earth. Aste-roids of the L chondrite type may normally rep-resent a smaller fraction of the total large-body£ux to Earth than the fraction of L chondritemeteorites in the total £ux of meteorite-sized ob-jects. For example, it is likely that the ratio ofcomet to asteroid particles impacting the top ofthe atmosphere increases with particle size, butcometary material may have a higher entry speed,causing the smaller comet objects to preferentiallydisintegrate to dust during atmospheric passage[38]. In essence, an order of magnitude increasein the in£ux of L chondritic asteroids following abreak-up event may not manifest itself clearly inthe imperfect cratering record. The distribution ofknown impact crater ages hence does not contra-dict a one to two orders of magnitude increase inmeteorite £ux as indicated by the fossil meteoritedata, particularly if the peak increase in meteorite£ux was restricted to a period of a few millionyears. Our search for fossil meteorites in con-densed limestone from other time intervals ofthe Ordovician has so far been unsuccesful, sug-gesting tentatively that the culmination of the me-

teorite in£ux coincided with the deposition of theOrthoceratite Limestone exposed in the Thorsbergquarry.

Acknowledgements

We thank the National Geographic Society, theSwedish Natural Science Research Council, theCrafoord Foundation, the Royal Society of Artsand Sciences in Go«teborg, the Go«teborg Univer-sity and the Woods Hole Oceanographic Institu-tion for ¢nancial support. We are grateful to G.Thor, S. Thor and S. Thor for pursuing the mete-orite search, A. Bevan for providing us with Nul-larbor meteorite data, D. Cornell for advice onchromite analyses, Lary Ball and the WHOIICP-MS facility for the use of their Finnigan El-ement in making Os isotope measurements and O.Gustafsson and T. Abbruzzese for technical assis-tance. Helpful reviews were provided by P. Bland,C. Koeberl and H. McSween.[AH]

References

[1] I. Halliday, A.T. Blackwell, A.A. Gri¤n, The £ux of me-teorites on the Earth's surface, Meteoritics 24 (1989) 173^178.

[2] P.A. Bland, Quanti¢cation of meteorite infall rates fromaccumulations in deserts, and meteorite accumulation onMars, in: B. Peucker-Ehrenbrink, B. Schmitz (Eds.), Ac-cretion of Extraterrestrial Matter Throughout Earth'sHistory, Kluwer Academics, New York, 2001, pp. 267^303.

[3] A.W.R. Bevan, P.A. Bland, A.J.T. Jull, Meteorite £ux onthe Nullarbor region, Australia, Geol. Soc. London Spec.Publ. 140 (1998) 59^73.

[4] B. Schmitz, M. Tassinari, Fossil meteorites, in: B. Peuck-er-Ehrenbrink, B. Schmitz (Eds.), Accretion of Extrater-restrial Matter Throughout Earth's History, Kluwer Aca-demics, New York, 2001, pp. 319^331.

[5] H.H. Nininger, Find a Falling Star, Paul S. Eriksson,New York, 1972, 254 pp.

[6] B. Schmitz, B. Peucker-Ehrenbrink, M. Lindstro«m, M.Tassinari, Accretion rates of meteorites and cosmic dustin the early Ordovician, Science 278 (1997) 88^90.

[7] M. Lindstro«m, Vom Anfang, Hochstand und Ende einesEpikontinentalmeeres, Geol. Rundsch. 60 (1971) 419^438.

[8] M. Lindstro«m, Diagenesis of Lower Ordovician hard-grounds in Sweden, Geol. Palaeontol. 13 (1979) 9^30.

[9] B. Schmitz, M. Lindstro«m, F. Asaro, M. Tassinari, Geo-

Fig. 8. Histogram of ages of all dated craters in crater com-pilation of [47] complemented by information on ages of twocraters (Lockne and Tva«ren) in [48,49]. Note the small peakin crater ages in 450^480 Ma bin.

EPSL 6057 13-12-01


chemistry of meteorite-rich marine limestone strata andfossil meteorites from the Lower Ordovician at Kinne-kulle, Sweden, Earth Planet. Sci. Lett. 145 (1996) 31^48.

[10] S. Bergstro«m, World-wide correlation of the base of theproposed Middle Ordovician global series: Conodont andgraptolite evidence, in: Geological Society of America Âbstracts with Programs, 2001, p. A-391.

[11] P. Thorslund, F.E. Wickman, J.O. Nystro«m, The Ordo-vician chondrite from Brun£o, central Sweden. I. Generaldescription and primary minerals, Lithos 17 (1984) 87^100.

[12] J.O. Nystro«m, M. Lindstro«m, F.E. Wickman, Discoveryof a second Ordovician meteorite using chromite as atracer, Nature 336 (1988) 572^574.

[13] E. Jarosewich, J.A. Nelen, J.A. Norberg, Reference sam-ples for electron microprobe analysis, Geostand. Newsl. 4(1980) 43^47.

[14] G. Ravizza, D. Pyle, PGE and Os isotopic analyses ofsingle sample aliquots with NiS ¢re assay preconcentra-tion, Chem. Geol. 141 (1997) 251^268.

[15] D. Hassler, B. Peucker-Ehrenbrink, G. Ravizza, Rapiddetermination of Os isotopic composition by spargingOsO4 into a magnetic-sector ICP-MS, Chem. Geol. 166(2000) 1^14.

[16] W.A. Cassidy, R.P. Harvey, Are there real di¡erencesbetween Antarctic ¢nds and modern falls meteorites?,Geochim. Cosmochim. Acta 55 (1991) 99^104.

[17] G.R. Huss, Meteorite mass distributions and di¡erencesbetween Antarctic and non-Antarctic meteorites, Geo-chim. Cosmochim. Acta 55 (1991) 105^111.

[18] T. Meisel, R.J. Walker, J.W. Morgan, The osmium iso-topic composition of the Earth's primitive upper mantle,Nature 383 (1996) 517^520.

[19] P. Ramdohr, The Opaque Minerals in Stony Meteorites,Elsevier, Amsterdam, 1973, 245 pp.

[20] K. Keil, On the phase composition of meteorites, J. Geo-phys. Res. 67 (1967) 4055^4061.

[21] A. El Goresy, in: D. Rumble III (Ed.) Reviews in Min-eralogy, Vol. 3, Oxide Minerals, Mineralogical Society ofAmerica, Washington, DC, 1976, pp. EG47ÊG72.

[22] F. Heide, F. Wlotzka, Meteorites ^ Messengers fromSpace, Springer, Berlin, 1995, 231 pp.

[23] A.E. Rubin, Mineralogy of meteorite groups, Meteorit.Planet. Sci. 32 (1997) 231^247.

[24] K.G. Snetsinger, K. Keil, T.E. Bunch, Chromite fromèquilibrated' chondrites, Am. Mineral. 52 (1967) 1322^1331.

[25] T.E. Bunch, K. Keil, K.G. Snetsinger, Chromite compo-sition in relation to chemistry and texture of ordinarychondrites, Geochim. Cosmochim. Acta 31 (1967) 1569^1582.

[26] T.E. Bunch, K. Keil, Chromite and ilmenite in non-chon-dritic meteorites, Am. Mineral. 56 (1971) 146^157.

[27] T.E. Bunch, K. Keil, Mineralogy and petrology of silicateinclusions in iron meteorites, Contrib. Mineral. Petrol. 25(1970) 297^340.

[28] L.H. Fuchs, E. Olsen, K.J. Jensen, Mineralogy, mineral-chemistry, and composition of the Murchison (C2) mete-orite, Smithson. Contrib. Earth Sci. 10 (1973) 1^39.

[29] C.A. Johnson, M. Prinz, Chromite and olivine in type IIchondrules in carbonaceous and ordinary chondrites: Im-plications for thermal histories and group di¡erences,Geochim. Cosmochim. Acta 55 (1991) 893^904.

[30] K. Keil, H. Haack, E.R.D. Scott, Catastrophic fragmen-tation of asteroids: evidence from meteorites, Planet.Space Sci. 42 (1994) 1109^1122.

[31] D.D. Bogard, Impact ages of meteorites: A synthesis,Meteoritics 30 (1995) 244^268.

[32] H. Haack, P. Farinella, E.R.D. Scott, K. Keil, Meteoritic,asteroidal, and theoretical constraints on the 500 Ma dis-ruption of the L chondrite parent body, Icarus 119 (1996)182^191.

[33] H. Craig, The origin of manganese nodules: a 120-yearwell-kept secret, EOS Trans. AGU 73 (1992) 255.

[34] Z. Jianhua, Middle Ordovician conodonts from the At-lantic faunal region and the evolution of key conodontgenera, Ph.D. Thesis, University of Stockholm, 1998,pp. 1^27.

[35] J. Villumsen, A.T. Nielsen, S. Stouge, The trilobite andconodont biostratigraphy of the upper Volkhov^lowerKunda deposits at Ha«llekis quarry, Va«stergo«tland, Swe-den, in: Working Group on the Ordovician Geology ofBaltoscandia ^ Abstract Volume, Geological Museum,Copenhagen, 2001, pp. 30^31.

[36] P.A. Bland, T.B. Smith, A.J.T. Jull, F.J. Berry, A.W.R.Bevan, S. Cloudt, C.T. Pillinger, The £ux of meteorites tothe Earth over the last 50,000 years, Mon. Not. R. As-tron. Soc. 283 (1996) 551^565.

[37] B. Lang, M. Kowalski, On the possible number and massof fragments from Pultusk meteorite shower, 1868, Mete-oritics 6 (1971) 149^158.

[38] D.E. Brownlee, The origin and properties of dust impact-ing the Earth, in: B. Peucker-Ehrenbrink, B. Schmitz(Eds.), Accretion of Extraterrestrial Matter ThroughoutEarth's History, Kluwer Academics, New York, 2001,pp. 1^12.

[39] G. Kurat, C. Koeberl, T. Presper, F. Brandsta«tter, M.Maurette, Petrology and geochemistry of Antarctic micro-meteorites, Geochim. Cosmochim. Acta 58 (1994) 3879^3904.

[40] D.E. Brownlee, B. Bates, L. Schramm, The elementalcomposition of stony cosmic spherules, Meteorit. Planet.Sci. 32 (1997) 157^175.

[41] A. Meibom, B.E. Clark, Evidence for the insigni¢cance ofordinary chondritic material in the asteroid belt, Meteorit.Planet. Sci. 34 (1999) 7^24.

[42] P. McConville, S. Kelley, G. Turner, Laser probe 40Ar^39Ar studies of the Peace River shocked L6 chondrite,Geochim. Cosmochim. Acta 52 (1988) 2487^2499.

[43] N. Nakamura, T. Fujiwara, S. Nohda, Young asteroidmelting event indicated by Rb^Sr dating of the Point ofRocks meteorite, Nature 345 (1990) 51^52.

[44] T. Fujiwara, N. Nakamura, Additional evidence of a

EPSL 6057 13-12-01


young impact-melting event on the L-chondrite parentbody, Lunar Planet. Sci. 23 (1992) 387^388.

[45] V. Zappala©, A. Cellino, B.J. Gladman, S. Manley, F. Mi-gliorini, Asteroid showers on Earth after family break-upevents, Icarus 134 (1998) 176^179.

[46] B.J. Gladman, F. Migliorini, A. Morbidelli, V. Zappala©,P. Michel, A. Cellino, C. Froeschle, H.F. Levison, M.Bailey, M. Duncan, Dynamical lifetimes of objects in-jected into asteroid belt resonances, Science 277 (1997)197^201.

[47] R.A.F. Grieve, The terrestrial cratering record, in: B.Peucker-Ehrenbrink, B. Schmitz (Eds.), Accretion of Ex-traterrestrial Matter Throughout Earth's History, KluwerAcademics, New York, 2001, pp. 379^402.

[48] Y. Grahn, J. No¬lvak, F. Paris, Precise chitinozoan datingof Ordovician impact events in Baltoscandia, J. Micropa-leontol. 15 (1996) 21^35.

[49] E. Sturkell, J. Ormo«, J. No¬lvak, Aî . Wallin, Distant ejectafrom the Lockne marine-target impact crater, Sweden,Meteorit. Planet. Sci. 35 (2000) 929^936.

EPSL 6057 13-12-01


A rain of ordinary chondritic meteorites in the early ... · orite of the project. From this bed...

Documents

Transcript of A rain of ordinary chondritic meteorites in the early ... · orite of the project. From this bed...