chicxulub impact ejecta from albion island,...

ELSEVIER Earth and Planetary Science Letters 170 (1999) 351–364www.elsevier.com/locate/epsl

Chicxulub impact ejecta from Albion Island, Belize

Kevin O. Pope a,Ł, Adriana C. Ocampo b, Alfred G. Fischer c, Walter Alvarez d, BruceW. Fouke e, Clyde L. Webster f, Francisco J. Vega g, Jan Smit h, A. Eugene Fritsche i,

Philippe Claeys j

a Geo Eco Arc Research, 3220 N Street NW, Suite 132, Washington, DC 20007, USAb Jet Propulsion Laboratory, California Institute of Technology, MS 183-601, 4800 Oak Grove Drive, Pasadena, CA 91109, USA

c Department of Earth Sciences, University of Southern California, University Park, Los Angeles, CA 90089, USAd Department of Geology and Geophysics, University of California, Berkeley, CA 94720, USA

e Department of Geology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USAf Geoscience Research Institute, Loma Linda University, Loma Linda, CA 92350, USA

g Instituto de Geologia, Universidad Nacional Autonoma de Mexico, Ciudad Universitaria, Mexico D.F., Mexicoh Institute of Earth Sciences, Vrije Universiteit, de Boelelaan 1085, 1081 HV Amsterdam, Netherlands

i Department of Geological Sciences, California State University, Northridge, CA 91330, USAj Museum fur Naturkunde, Institut fur Mineralogie, D-10115 Berlin, Germany

Received 21 January 1999; accepted 14 May 1999

Abstract

Impact ejecta from the Albion Formation are exposed in northern Belize. The ejecta come from the outer portion of thecontinuous ejecta blanket of the Chicxulub crater, which is located 360 km to the northwest. The basal unit of the AlbionFormation is a ¾1-m-thick clay and dolomite spheroid bed composed of up to four discrete flows. The clay spheroids arealtered impact glass, and the dolomite spheroids are accretionary lapilli. The upper unit is a ¾15-m-thick coarse diamictitebed containing altered glass, large accretionary blocks, striated, polished, and impacted cobbles, and rare shocked quartz.The abundance of accretionary clasts, evidence for atmospheric drag sorting, and the presence of multiple flows in theAlbion Formation indicate that atmospheres play an important role in the formation of the outer portions of continuousejecta blankets of large craters. 1999 Elsevier Science B.V. All rights reserved.

Keywords: ejecta; impact features; Chicxulub Crater; Belize

1. Introduction

The Albion Formation was first described byOcampo et al. [1] in a quarry on Albion Island lo-cated in the Hondo River of northern Belize (Fig. 1).Sedimentologic and petrographic evidence suggest

Ł Corresponding author. Tel.=Fax: C1-202-965-5056; E-mail:[email protected]

that the Albion Formation is proximal impact ejectaand its stratigraphic position, overlying Late Creta-ceous platform carbonates, indicates that the ejectamost likely come from the Chicxulub impact crater,which lies approximately 360 km northwest of Al-bion Island [1]. The Albion quarry, together witha series of Albion Formation outcrops that we dis-covered on the Mexican side of the Hondo River inJanuary 1998, are the closest ejecta exposures yet

0012-821X/99/$ – see front matter 1999 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 9 9 ) 0 0 1 2 3 - 5

352 K.O. Pope et al. / Earth and Planetary Science Letters 170 (1999) 351–364

Fig. 1. Stratigraphy of the Albion Island quarry, Belize. Spheroid bed section is from the saddle location; strata 1–4 are described intext, substrata (e.g. 2a, 2b) are subdivisions not found in every quarry location. Clasts in diamictite bed are schematic. Key to BartonCreek Formation dolomite: a D thick, coarsely crystalline beds with vugs indicative of leached anhydrite nodules; b D thin, finelycrystalline beds; c D cross-bedded, coarse, calcarenite beds; and d D fine-grained, organic-rich, laminated beds with hummocky andripple cross-bedding. Inset: map with location of Albion Island and the Chicxulub crater.

found to the crater and comprise the only known ex-posures of the Chicxulub continuous ejecta blanket.In this paper we expand upon this earlier work andreport new findings from the ongoing research onAlbion Island.

2. Barton Creek Formation

The Barton Creek Formation on Albion Island iscomposed of a tan-colored, 26-m-thick sequence ofdolomitized limestone deposited in a shallow-water

platform environment (Fig. 1). Petrographic and out-crop analyses indicate interbedding of shallow sub-tidal and intertidal sediments, some with nodularmolds suggesting periodic emergence and evaporitedeposition. Cathodoluminescence (CL) petrographyindicates a complex paragenetic sequence of calcitecementation and dolomitization [2]. Benthic forami-nifera fossils are present in the section, but they aretoo recrystallized to be identified. A new species offossil crab, Carcineretes planetarius [3] (Fig. 2a),and nerineid gastropods (Fig. 2b) are found nearthe base of the exposed section (Fig. 1). Flores [4]

K.O. Pope et al. / Earth and Planetary Science Letters 170 (1999) 351–364 353

Fig. 2. Barton Creek Formation: (a) Carcineretes planetarius crab (scale bar 10 mm); and (b) nerineid gastropod (scale bar 1 mm). SeeFig. 1 for location in section.

interpreted the Barton Creek Formation as represent-ing a shallow, back-reef lagoon environment. This isconsistent with the gastropods and crabs, as well aswith the overall restricted fauna and lithology of theAlbion quarry.

The Barton Creek Formation in central Belizecontains Upper Cretaceous fossil rudistids, mili-olids, rotalids, Dyciclina sp., Lockhartia sp., Num-moloculina sp., and Valvulina sp. [4]. The ner-ineid gastropods from Albion Island have infold-ing wall structures (Fig. 2b) characteristic of thelast developmental stage, indicating a Campanianto Maastrichtian age. Vega et al. [3] propose aMaastrichtian age for the Carcineretes planetariuscrabs found at Albion Island. The only other mem-ber of Carcineretes, C. woolacotti, comes from theearly Maastrichtian [5]. These findings, together withthe 87Sr=86Sr isotope ratios (0.70786–0.70796) fromBarton Creek Formation reported by Ocampo et al.[1], suggest a Maastrichtian age for the dolomitizedlimestones underlying the impact ejecta.

The Barton Creek Formation in the quarry wasfolded into a gently plunging anticline, the apexof which was eroded prior to deposition of theoverlying Albion Formation. The upper surface ofthe Barton Creek Formation is irregular with localabrupt relief of 20–50 cm and preserved patches ofcaliche (Fig. 3). This caliche is composed of poorlysorted, angular fragments of the Barton Creek For-mation dolomite re-cemented with micritic dolomite

and coarse-grained calcite with iron oxide crusts.Such a surface is similar to the caliche caps found onemergent portions of the Pleistocene Yucatan plat-form [6]. These findings indicate that Albion Islandwas emergent prior to the deposition of the AlbionFormation.

3. Albion Formation spheroid bed

The spheroid bed ranges in thickness from 0.10to 1.72 m, but most exposures are ¾1 m thick.The basal, and occasionally the upper, contact isa ¾2-cm-thick brown clay layer (Fig. 4) withpronounced slickensides. Similar clay layers occuralong shear planes that cut diagonally across the bed(Fig. 5). Typically, the upper contact is marked bythe abrupt appearance of cobbles of the overlyingdiamictite bed, but with little change in the matrix(Fig. 5). Four distinct strata with gradational contactsare differentiated by clast type and foliation (Fig. 1).All four strata contain 10–30% (by volume) dolomitespheroids, 1–10 mm in diameter. They also contain10% (by volume) flakes and angular clasts of brownclay, 1–3 mm in diameter. Spheroids are supportedin a weakly consolidated, fine-grained, dolomite andclay matrix. Rare dolomite spheroids reach sizes of20–30 mm. Green clay spheroids 1–10 mm in diam-eter comprise 10–30% (by volume) of stratum 1 andgrade upward into the base of stratum 2 (2a, Fig. 1).


Fig. 3. Upper contact of the Barton Creek Formation. Note caliche with brecciated surface and dark iron oxide staining. Pen is 14 cmlong and rests on land surface at time of impact.

Most spheroids fall in the 5–10 mm size range. Nosize grading of spheroids is apparent in outcrop.

Stratum 1 is 15–20 cm thick. The dolomite andgreen clay spheroids are flattened (Fig. 6) and de-posited with the long axis parallel to the underlyingBarton Creek Formation surface (imbricated), giv-

Fig. 4. Albion Island quarry. Near center of photograph is a planar 80-cm-thick Albion Formation spheroid bed exposure with sharplower (Barton Creek Formation) and upper (Albion Formation diamictite bed) contacts.

ing the bed a foliated appearance, which followsthe microtopography. Stratum 2 is 25–40 cm thick,contains slightly flattened dolomite spheroids, andhas faint foliation. Stratum 3 is 40–50 cm thickand contains spherical to slightly oblate dolomitespheroids, distinct foliation, and 10% (by volume)


Fig. 5. Albion Formation spheroid bed with clay lined shearplane cutting diagonally across the bed near the base. Scalewith 10-cm increments rests on Barton Creek Formation contact(total exposure 2 m). Note contact between spheroid bed anddiamictite bed at 70 cm, marked by the appearance of cobblesbut little change in the matrix.

white, chalky, angular dolomite clasts in the samesize distribution as the spheroids. Stratum 4 is 40–50cm thick and has the same composition as stratum 3,but no foliation.

3.1. Matrix and dolomite spheroids

The matrix of the spheroid bed is composed ofsucrosic 25–50 µm dolomite crystals with concen-tric red=yellow CL zonations, which contrast withthe bright, concentric red=orange CL zonations inthe dolomite crystals of the spheroids [2]. The ma-trix and spheroids contain interstitial iron oxide and

Fig. 6. Dolomite spheroids from the base of the Albion For-mation spheroid bed. Note flattened disk shape (scale bar 10mm).

smectite. The smaller (<5 mm) dolomite spheroidsare composed of crystals of uniform size (¾25 µm).Most dolomite spheroids in the 10–25 mm rangehave concentric bands of fine and coarse crystalsand iron oxide staining surrounding an angular core(Fig. 7). Coarse spheroid bands have crystal sizes upto 500 µm, but most bands have 50–75 µm crystals.Many of the larger crystals within these bands haveblack opaque cores (Fig. 7c). One large dolomitespheroid contains a core composed of two smallerspheroids joined together (Fig. 8). The angular coresare composed of a white, chalky, fine-grained (¾10µm) dolomite (Fig. 7a) that is similar to the dolomiteclasts in the upper strata of the spheroid bed.

3.2. Clay spheroids

The green clay spheroids at the base of thespheroid bed have a dull, waxy appearance. X-raydiffraction (XRD) analyses indicate that they aremostly smectite. Petrographic and microprobe anal-yses reveal that portions of some spherules areisotropic and have Si=Al ratios (Table 1) higherthan typical smectites, as found in other K=T bound-ary clay spherules [7,8]. Microprobe analyses ofmajor oxides indicate that the green clay spheroidsare compositionally similar to the Haitian palago-nites formed from K=T boundary glass spherules


Fig. 7. Dolomite spheroid (accretionary lapilli) from the Albion Formation spheroid bed: (a) cross-section showing white angular coreand concentric bands (scale bar 10 mm); (b) thin section (plain polarized light) showing fine-grained core (dark in transmitted light) withconcentric bands of coarse and fine grain dolomite (scale bar 5 mm); (c) thin section (plain polarized light) of concentric band of coarsecrystals showing common, black, opaque cores (scale bar 0.5 mm).

Table 1Major oxide analyses from the Albion Formation diamictite bed and spheroid bed, and from K=T boundary in Haiti

Spheroid a Diamictite a Spheroid b Haiti K=T a

(n D 12) (n D 11) (n D 3) (n D 5)

Na2O 0.06 š 0.02 0.06 š 0.05 0.04 š 0.02 0.6 š 0.2MgO 9.65 š 0.72 8.56 š 0.51 17.60 š 1.91 6.5 š 0.4Al2O3 16.70 š 1.17 16.42 š 0.67 3.69 š 2.17 15.7 š 1.6SiO2 58.21 š 3.27 57.33 š 2.31 11.05 š 6.70 61.1 š 1.8K2O 0.36 š 0.07 2.00 š 1.06 0.13 š 0.14 0.2 š 0.1CaO 0.13 š 0.22 0.13 š 0.04 24.33 š 4.35 0.7 š 0.1TiO2 0.65 š 1.13 0.93 š 0.78 0.17 š 0.09 0.5 š 0.2MnO 0.02 š 0.01 0.01 š 0.01 <0.01 no dataFeO 3.43 š 0.50 6.02 š 2.09 1.53 š 0.49 7.6 š 1.0

L.O.I. 41.57 š 3.39

Total 89.20 š 4.95 91.46 š 2.25 100.13 š 0.15 92.9 š 3.5

L.O.I. D loss on ignition.a Microprobe analyses of palagonite shards selected under SEM, error is 1 standard deviation. Haiti spherule analyses from Bohor andGlass [7].b XRF analyses of oven-dried bulk samples.


Fig. 8. Thin section (plain polarized light) of dolomite spheroid from the Albion Formation spheroid bed with a core of two highlyporous dolomite spheroids (scale bar 10 mm).

(Table 1). The typical size (mostly 5–10 mm in di-ameter) of the Albion Formation spheroids is largerthan that of the Haitian spherules (mostly <2 mmin diameter [7,9]). Scanning electron micrographs(SEM) and thin sections of the clay spheroids showvesicular and spherulitic textures typical of devitri-fied glass (Fig. 9).

4. Albion Formation diamictite bed

Ocampo et al. [1] estimated the maximum thick-ness of the diamictite to be about 15 m, but anerosion surface cuts the top and its original thicknessis unknown. The diamictite is weakly consolidated,unstratified, poorly sorted, and is composed mostlyof matrix-supported carbonate and clay clasts. Car-bonate clasts are angular to subrounded and range insize from sub-millimeter grains to boulders about 8m in diameter.

4.1. Carbonate clasts and matrix

Due to the abundance of large clasts in the di-amictite, we employed a grain size analysis tech-

Fig. 9. SEM image of a polished section of a palagonite fragmentfrom a spheroid bed green clay spheroid showing a relict vesicle(right) and spherulitic devitrification features (lower left) (scalebar 0.1 mm).

nique using the cross-sectional area of clasts, similarto that used in thin section analysis. We measuredclasts diameters within 1 m by 1 m grids at fouroutcrops in the quarry, and in photographs of a 66-m-long freshly cut face of the quarry. The outcropswere randomly chosen from the base (saddle and


Fig. 10. Histograms of Albion Formation diamictite bed grainsize measured in 1 m by 1 m grids from the saddle (SAD), cut1 lower part (CTL), cut 1 upper part (CTU), and cut 2 (CT2)locations in quarry. Percentages are based on cross-sectional areawithin the grids. Matrix is all clasts smaller than 0.8 cm. Lineon graph depicts mean grain size for the four locations. Phi isa base-2 scale with phi �4 D 0.8–1.6 cm and a phi �8 D12.8–25.6 cm.

cut 1 lower part) and upper (cut 1 upper part, cut2) portions of the diamictite bed. We measured allclasts larger than 0.8 cm within the grids and largerthan 2 m in the photographs (smaller clasts werenot detected in the photographs). The percent ma-trix (defined here as clasts smaller than 0.8 cm indiameter) was calculated for each grid. The resultsof these analyses are shown in a series of histogramsfor the grid data (Fig. 10) and as a graph of the cu-mulative percentage (area) for the combined grid andphotograph-based data (Fig. 11). Assuming uniformdensity of materials, these data are good proxies forweight percentages more commonly used in suchanalyses. These figures illustrate the poorly sortednature and heterogeneity of the diamictite bed.

Most of the carbonate clasts are dolomitized lime-stone and the nonclastic component of the matrix isa micritic dolomite (10–60 µm crystals) with minoramounts of interstitial clay and iron oxide. Multi-ple generations of CL-distinct dolomite cements arefound in the matrix and in the carbonate clasts [2].The diamictite bed contains rare concentric dolomitespheroids 10–20 mm in diameter and a few largeblocks (up to 4.5 m in diameter) with lithic cores and7–75 cm thick rinds of material similar in textureand composition to the diamictite matrix (Fig. 12).Rare carbonate clasts exhibit striations in a single or

Fig. 11. Albion Formation diamictite bed grain size plotted ascumulative percent (area) on a probability scale. Data in phi �4to �8 are mean values from grid analysis (Fig. 10). Data forphi �12 and �13 are from analysis of photographs. Note griddata plot near a straight line, indicative of a single normallydistributed population.

Fig. 12. Albion Formation diamictite bed accretionary block(diameter 4.5 m); dashed lines demarcate greenish gray core ofdolomite and clay surrounded by a 75-cm-thick rind of whitedolomite matrix.


Fig. 13. Albion Formation diamictite bed (a) striated dolomitecobble with close-up of upper tip (b) showing penetrating lithicfragment (scale bar in both photographs 10 mm).

occasionally multiple directions and in many casesterminate in an abrupt angular surface (Fig. 13a).These striations range from microscopic sets thatgrade into a fine polish, to parallel grooves 1–2mm deep. Sand-sized carbonate grains occur withinsome striae. In one sample, a dark dolomite rock

Fig. 14. Albion Formation diamictite bed calcite spherules. Noteprotrusions on surface (scale bar 10 mm).

chip penetrated the clast to form a pit with a striatedwall (Fig. 13b). Ocampo et al. [1] reported find-ing a clay pocket filled with radially fibrous calcitespherules that they called pisoids. We found anothersimilar pocket in 1995. These new radially fibrouscalcite spherules included many with angular coresof a dark gray dolomite. The calcite spherules haveetched surfaces and wart-like protrusions (Fig. 14)similar to the smaller altered glass spherules fromthe K=T boundary [9].

4.2. Clay clasts and other components

Green clay clasts comprise about 10% (by weight)of the diamictite bed. They range in size from <1mm to 40 mm, but most are <10 mm in diameter.Petrographic, XRD, SEM, and microprobe analysesof the clays reveal that they are smectite and palag-onite. Compositionally, these palagonites are similarto those of the spheroid bed, except that they con-tain slightly more K and Fe (Table 1). Many of theclay fragments have the same vesicular textures andspherulitic devitrification features found in the greenclay spheroids of the spheroid bed (Fig. 15).

Authigenic quartz found within vugs in the di-amictite bed contains anhydrite inclusions. Acidleach residues from the diamictite contain commongypsum as finely disseminated crystals. Apparently,sulfates have been mostly leached from the deposits.We found a single detrital quartz grain in the acidleach residues that has one prominent set and a sec-ond clear set of planar deformation features (Fig. 16),indicative of shock. Detrital quartz grains were ex-tremely rare in the over 1 kg of diamictite that wasprocessed and only one shocked quartz grain wasfound.


Fig. 15. Thin section (plain polarized light) of Albion Forma-tion diamictite bed altered glass shard (palagonite and smectite)showing relict vesicles with concentric alteration patterns (scalebar 0.5 mm).

Fig. 16. Albion Formation diamictite bed shocked quartz grainshowing two sets of planar deformation features (scale bar 20µm).

5. Interpretations

The shocked quartz and impact-penetrated cob-bles add to evidence provided by Ocampo et al.[1] that the Albion Formation is of impact ori-gin. The penetrating rock chip must have impactedunder pressures and=or temperatures sufficient toplastically deform carbonates [10], since impacts be-tween solid particles produce radial fractures andspallation zones [11], not the grooved penetrat-ing features we observed. The origin of this hightemperature=pressure environment is uncertain, butmay have developed during particle ejection from thecrater or within the ejecta curtain. The impact-pene-trating chip is part of a cluster of pits and striationsthat radiate away from the tip of the pointed cob-ble in parallel sets (Fig. 13). This suggests that thecobble was rotating as it past through (in a relativesense) a cloud of finer debris. Similar features arefound in other ejecta [12] and these surface fea-tures are common on pebbles and cobbles in theChicxulub ejecta deposits in central Belize [13]. In-dications that the Barton Creek Formation underly-ing the Albion Formation is probably Maastrichtian,provides an improved temporal link between theAlbion Formation and the nearby K=T Chicxulubcrater.

The Barton Creek Formation carbonates are com-posed of several size-distinct and CL-distinct gener-ations of cement and replacement dolomite, whichare different with respect to size and CL from thedolomites in either the spheroid or the diamictitebeds. Furthermore, the dolomites in the spheroid bedare different in size, CL character, and composi-tion from the dolomites comprising the matrix andallochthonous blocks of the diamictite bed. No ev-idence of dolomite cement overgrowths on detritaldolomite crystals or fragments was observed withCL in either the spheroid or the diamictite beds.These observations suggest that: (1) the fluids thatdolomitized the Barton Creek Formation, spheroidbed, and diamictite bed were chemically distinctfrom one another; (2) the Barton Creek Formationlimestone was dolomitized prior to deposition of theejecta; and (3) neither the spheroid bed dolomitesnor most of the dolomitized allochthonous blocksin the diamictite bed were derived from the directlyunderlying Barton Creek Formation.


Fig. 11 indicates that the diamictite contains twopopulations of clasts. The grid data represent onenormally distributed population with a modal size(excluding matrix) of �5 to �6 phi (3.2–6.4 cm,Fig. 10), which plots close to a straight line on theprobability scale of the cumulative % curve (Fig. 11).The photographic data, representing the larger clastsizes (�11 to �13 phi, ¾2–8 m), do not plot closeto the straight line of the grid data. Such a devia-tion from a single straight line is a good indicationthat two populations are involved [14]. While it ispossible that the different measurement techniqueshave some effect on the results, we note that theproposed difference between the two populations isthat large clasts are too abundant to be the tail of asingle population distribution. One would expect thatif the photographic technique did introduce someerror, it would underestimate the number of largeclasts. The lack of clasts in the �8 to �11 phi sizerange (¾25–200 cm) may in part be an artifact of thesampling, since it is difficult to detect clasts in thisrange in the photographs. Nevertheless, we empha-size that no clasts in this size range were found inthe randomly placed grids. Furthermore, we note thatclasts this size were very rarely encountered duringthe field work and there may be a true gap in clastsize between the two proposed populations.

Ocampo et al. [1] interpreted the green clayspheroids and clasts to be altered impact glass andour additional analyses support this. The microprobeanalyses indicate that no glass is preserved, but muchof the clay is palagonite, a common alteration prod-uct of glass. The carbonate spheroids found in thespheroid bed and rarely in the diamictite bed areinterpreted to be accretionary lapilli, which are com-monly found in pyroclastic deposits. Ocampo et al.[1] proposed an accretionary lapilli origin for thespheroids with cores, but suggested other possibleorigins for the spheroids without cores. Pyroclasticdeposits commonly contain lapilli with and with-out cores [15] and our petrographic examination ofthe carbonate spheroids found no significant differ-ences in the matrix composition between spheroidswith and without cores. The radially fibrous calcitespherules found in the diamictite bed may be an-other form of accretionary lapilli, given that many ofthese also have lithic cores. Their fibrous crystallinestructure and tektite-like surfaces noted above indi-

cate that these spherules may have crystallized fromcalcite melt.

The complex stratification within the spheroid bedindicates that the composition of the volatile-richplume that accreted the lapilli was rapidly evolv-ing as it reached the southern end of the YucatanPeninsula. The several discrete, thin strata suggestthat multiple flows with different amounts of glass,lapilli, and carbonate clasts were involved. The ab-sence of graded beds and the common lack of shearzones between strata within the spheroid bed sug-gest that the debris may have been in part depositedby turbulent flows. Pyroclastic air fall deposits areusually normally graded and laminar flows produceinverse grading due to the expulsion of large clastsfrom the base by shear forces [16]. Nevertheless, thefoliation (imbrication) found in the spheroid bed iscommonly observed in pyroclastic flows, where it isattributed to shear during laminar flow [17]. We alsonote that a few spheroid bed exposures have shearplanes (Figs. 4 and 5). The presence of foliation andbasal shear in some strata, and the lack of thesefeatures in others, may indicate that the flows thatdeposited the spheroid bed were changing betweenturbulent and laminar. Changes from a turbulent tolaminar flow regime are typical of pyroclastic flowswhere the loss of volatiles (deflation) reduces buoy-ancy forces [16]. Such flow regime changes are typi-cal of distal pyroclastic flows shortly before deflationand friction halt the flow [18]. This is consistent withour geological reconnaissance, which indicates thatAlbion Island is near the terminus of the continuousejecta blanket.

We interpret the large blocks with coatings ofmatrix material in the diamictite bed as accretionaryblocks similar in origin to the carbonate spheroids.Ocampo et al. [1] originally described one of theselarge (3 m) clasts from Albion Island as a mud-coated boulder (fig. 11 in [1]). They suggested thatthe mud coating may be genetically related to alarge mudball also found in the diamictite bed (fig.14 in [1]), which was interpreted as being similarin origin to mudballs in terrestrial debris flows. In1998 a better exposure (due to erosion) of this mud-ball revealed that it is a giant mud-coated boulder(Fig. 12), not a mudball. Mudballs form in debrisflows when fine-grained sediment (mud) is rippedup and carried along in the flow [19]. In contrast,


mud-coated boulders like those found on Albion Is-land have not been reported from debris flows. Wenow recognize that these large mud-coated bouldersare similar in structure and composition to the ac-cretionary lapilli, hence our interpretation of them asaccretionary blocks. These large accretionary blocksindicate that the diamictite bed is also derived fromdebris once suspended in a volatile-rich plume, butone with large clasts and sufficient energy to trans-port 4.5-m-diameter boulders.

6. Comparisons with other ejecta deposits

Surface exposures of proximal ejecta from largeimpact craters are rare on Earth. Prior to the Be-lize discoveries, the best-known examples were fromthe Ries crater in Germany [20,21] and the Sud-bury crater in Canada (e.g. the Onaping Formation,[22]). The Alamo breccia in Nevada [23] is prob-ably proximal ejecta from a large impact, but nocrater has been identified. The Pelarda Formation inSpain may also be proximal impact ejecta from theAzuara structure [24,25], but the impact origin ofthis material is controversial. Ries (¾24 km diame-ter) is much smaller than Chicxulub. The exposuresof Ries ejecta extend about two crater radii (24 km)from the center of the crater, whereas Albion Islandis over three crater radii (360 km) from the centerof the Chicxulub. Thus, Albion Island represents amore distal environment in both relative and absoluteterms. Given these contrasts in size and proximity,depositional processes may have been different forthe Ries and Albion Island deposits. Sudbury (¾200km) is approximately the same size as Chicxulub,but no ejecta from outside the crater is preserved.Ejecta from other large craters is either poorly pre-served (e.g. Popagai crater) or is known only fromdrilling (e.g. Chesapeake Bay crater). The AlbionFormation is the only known well-exposed exampleof the continuous ejecta blanket from a large (>50km diameter) crater. Chicxulub ejecta from farthersouth in Belize [10] are probably not part of thecontinuous ejecta blanket.

There are no good impact analogues to the Al-bion Formation spheroid bed. The best analogy is avolcanic pyroclastic flow; however, we know of novolcanic examples of carbonate accretionary lapilli

or blocks. Carbonate accretionary lapilli are reportedfrom the Alamo breccia [26] and from Azuara [25].They are also found in more distal Chicxulub ejectain Mexico [27,28]. Silicate accretionary lapilli arepresent in crater fill breccias of the Sudbury [22]and Manson [29] craters. The melt-rich ejecta calledsuevite at Ries crater contain silicate spherules, chon-drules, and accretionary lapilli [30]. In none of theseexamples do lapilli form a distinct stratum beneathcoarser ejecta, and in fact, the stratigraphy at Riescrater appears reversed, where the lapilli-bearingstrata of the suevite overlie the coarse Bunte breccia.

The Bunte breccia of the Ries crater and theBunte-like breccia drilled on the Chicxulub rim [31]share several characteristics with the Albion For-mation diamictite bed [1]. It is important to note,however, that the Bunte breccia at Ries containsno altered glass or accretionary clasts. The AlbionFormation diamictite bed contains abundant alteredglass and accretionary clasts, and thus appears to bea mixture of Bunte breccia and suevite-like compo-nents. The diamictite bed also shares many charac-teristic with the Onaping Formation of the Sudburycrater [22], notably common altered glass fragmentsand accretionary lapilli and blocks. We also noteclose similarities between the clay pockets with ra-dially fibrous calcite spherules in the diamictite bedand some of the ‘spherulitic dikes’ (those that are ir-regularly shaped bodies, not dikes) found in the On-aping Formation, which contain cm-sized spheruleswith radiating feldspar crystals [22].

7. Continuous ejecta blanket deposition

The widely accepted model for emplacement ofthe Bunte breccia at Ries crater is ballistic sedi-mentation [32], where impacting ejecta blocks scourthe surface and produce a surface flow composedof minor amounts of ejecta and much local rock.The principal evidence for this model is the abun-dance of local rock in the Bunte breccia [21]. Mostof the lithologies in the Albion Formation are sim-ilar to those of the local Barton Creek Formationand could be the product of scouring by secondaryimpacts on the Yucatan platform. Nevertheless, thisargument is not conclusive, since the upper 2 km oftarget rock at Chicxulub is also composed of car-


bonates similar to the Barton Creek Formation [33].Impact models [34] predict that the high ejectionvelocities needed to reach Belize would only occurat shallow target depths, or close to the impact cen-ter where the deeper crystalline rocks would havebeen melted (origin of the Albion Formation alteredglass). Therefore, we cannot rule out the possibil-ity that most of the Albion Formation is composedof primary ejecta from Chicxulub. We note that themore distal ejecta from Ries are also dominated bythe uppermost target lithologies [21].

The fact that there is no evidence for major scour-ing at the base of the Albion Formation, and that theCL studies indicate that the diamictite bed and Bar-ton Creek Formation dolomites are different, arguesagainst production by local secondary impacts. Nev-ertheless, the ejecta may have been produced by suchimpacts farther up range. More importantly, ballisticsedimentation does not account for the abundance ofaccretionary lapilli and blocks in the Albion Forma-tion.

Another model of continuous ejecta blanket em-placement is the ring vortex model [35–37]. In thismodel a turbulent cloud of primary ejecta is pro-duced by atmospheric drag-induced ring vorticesthat develop upon breakdown of the advancing ejectacurtain. The turbulent cloud thus produced containsmuch of the finer ejecta, which is stripped from theejecta curtain. This model provides a better explana-tion of the Albion Formation, given the abundanceof fine matrix and accretionary and rounded clasts(abrasion in the ring vortices?). The large accre-tionary blocks in the Albion Formation may attest tothe power of this turbulence, which likely reachedsupersonic speeds [35,36]. The apparent sorting re-flected in the two populations of clasts in the diamic-tite bed may also be a product of ejecta interactionwith the atmosphere. The ring vortex model [36]predicts sorting of clasts in the same size range asthat found in Belize.

The complex stratigraphy of the Albion Forma-tion indicates that the ejecta were emplaced by mul-tiple flows of different composition and possiblyorigin. The origin of the spheroid bed debris maybe either jetting or vapor plume entrainment of finematerial from the ejecta curtain. The diamictite bedis probably derived from collapse of the main ejectacurtain. These flows may be the product of the com-

plex emplacement processes proposed for fluidizedejecta blankets on Mars and Venus [35,37].

We propose that ballistic sedimentation on plan-ets with atmospheres may dominate only near thecrater rim (perhaps within a final crater radii ofthe rim), where the initial blast of the vapor plumeand the advancing ejecta curtain largely displace theatmosphere, and where permeability in the ejectacurtain has not yet developed. At greater distances,the effects of the atmosphere become significant,secondary impacts are reduced, and ejecta deposi-tion is dominated by material derived from turbulentclouds of debris. This study demonstrates that atmo-spheres can play an important role in the formationof continuous ejecta blankets of large craters.

Acknowledgements

This work was funded by the Exobiology Programof the National Aeronautics and Space Administra-tion and by The Planetary Society. We thank ThePlanetary Society volunteers and the Albion StoneCompany for their help with the fieldwork, and theBelize Geology and Petroleum Office for their per-mission to conduct the research. [MK]

References

[1] A.C. Ocampo, K.O. Pope, A.G. Fischer, Ejecta blanketdeposits of the Chicxulub crater from Albion Island, Be-lize, in: G. Ryder, D. Fastovsky, S. Gartner (Eds.), TheCretaceous–Tertiary Event and Other Catastrophes in EarthHistory, Geol. Soc. Am. Spec. Pap. 307 (1996) 75–88.

[2] B.W. Fouke, W. Alvarez, Ph. Claeys, A. Ocampo, K.O.Pope, J. Smit, F.J. Vega, Cathodoluminescence study ofcarbonate growth phases in bedrock and KT ejecta fromthe Chicxulub Impact, at the Albion Island quarry, Belize,Geol. Soc. Am. Abstr. Progr. 28 (1996) A-183.

[3] F.J. Vega, R.M. Feldman, A.C. Ocampo, K.O. Pope, A newspecies of Late Cretaceous carcineretid crab (Brachyura:Carcineretidae) from Albion Island, Belize, J. Paleontol. 71(1997) 615–620.

[4] G. Flores, Geology of northern British Honduras, Am.Assoc. Pet. Geol. Bull. 36 (1952) 404–409.

[5] S.F. Morris, The fossil arthropods of Jamaica, in: E.M.Wright, E. Robinson (Eds.), Biostratigraphy of Jamaica,Geol. Soc. Am. Mem. 182 (1993) 115–124.

[6] W.C. Ward, Quaternary geology of northeastern YucatanPeninsula, in: W.C. Ward, A.E. Weidie, W. Back (Eds.),


Geology and Hydrogeology of the Yucatan and QuaternaryGeology of Northeastern Yucatan Peninsula, New OrleansGeological Society Publication, 1985, pp. 23–95.

[7] B.F. Bohor, B.P. Glass, Origin and diagenesis of K=T im-pact spherules — from Haiti to Wyoming and beyond,Meteoritics 30 (1995) 182–198.

[8] D.A. Kring, W.V. Boynton, Altered spherules of impactmelt and associated relic glass from the K=T boundarysediments in Haiti, Geochim. Cosmochim. Acta 55 (1991)1737–1742.

[9] G.A. Izett, Tektites in Cretaceous=Tertiary boundary rockson Haiti and their bearing on the Alvarez impact extinctionhypothesis, J. Geophys. Res. 96 (1991) 20879–20905.

[10] J.R. Marshall, C. Bratton, K.O. Pope, A.C. Ocampo, Di-agnostic clast-texture criteria for recognition of impact de-posits (abstr.), Lunar Planet. Sci. 29 (1998) 1134.

[11] M.S. Prasad, M. Sudhakar, Collisions in the ejecta plume ofthe Australasian impact event, Lunar Planet. Sci. 27 (1996)1053–1054.

[12] E.C.T. Chao, Mineral produced high pressure striae andclay polish: key evidence for nonballistic transport of ejectafrom Ries crater, Science 194 (1976) 615–618.

[13] A.C. Ocampo, K.O. Pope, A.G. Fischer, Carbonate ejectafrom the Chicxulub crater: evidence for ablation and parti-cle interaction under high temperatures and pressures (ab-str.), Lunar Planet. Sci. 28 (1997) 1861.

[14] G.S. Visher, Grain size distribution and depositional pro-cesses, J. Sediment. Petrol. 39 (1969) 1074–1106.

[15] R. Schumacher, H.-U. Schmincke, Internal structure andoccurrence of accretionary lapilli — a case study at LaacherSee Volcano, Bull. Volcanol. 53 (1991) 612–634.

[16] M.F. Sheridan, Emplacement of pyroclastic flows: a review,in: C.E. Chapin, W.E. Elston (Eds.), Ash-Flow Tuffs, Geol.Soc. Am. Spec. Pap. 180 (1979) 125–136.

[17] C.E. Chapin, G.R. Lowell, Primary and secondary flowstructures in ash-flow tuffs of the Gribbles Run paleoval-ley, central Colorado, in: C.E. Chapin, W.E. Elston (Eds.),Ash-Flow Tuffs, Geol. Soc. Am. Spec. Pap. 180 (1979)137–154.

[18] K.H. Wohletz, M.F. Sheridan, A model of pyroclastic surge,in: C.E. Chapin, W.E. Elston (Eds.), Ash-Flow Tuffs, Geol.Soc. Am. Spec. Pap. 180 (1979) 177–194.

[19] W.B. Bull, Alluvial fans and near-surface subsidence inwestern Fresno County, California, U.S. Geol. Surv. Prof.Pap. 437-A (1964) 70 pp.

[20] E.C.T. Chao, R. Huttner, H. Schmidt-Kaler, Principal ex-posures of the Ries meteorite crater in southern Germany,Bayerisches Geologisches Landesamt, Munchen, 1978, 84pp.

[21] F. Horz, R. Ostertag, D.A. Rainey, Bunte Breccia of theRies: continuous deposits of large impact craters, Rev.Geophys. Space Phys. 21 (1983) 1667–1725.

[22] T.L. Muir, W.V. Peredery, The Onaping Formation, in: E.G.Pye, A.J. Naldrett, P.E. Giblin (Eds.), The Geology and OreDeposits of the Sudbury Structure, Ontario Geol. Soc. Spec.Vol. 1 (1984) 139–204.

[23] J.E. Warme, C.A. Sandberg, Alamo Megabreccia: record ofa Late Devonian impact in southern Nevada, GSA Today 6(1996) 1–7.

[24] K. Ernston, F. Claudin, Pelarda Formation (Eastern IberianChains, NE Spain): ejecta of the Azuara impact structure,Neues Jahrb. Geol. Palaeontol. Monatsh. H10 (1990) 581–599.

[25] K. Ernston, J. Fiebag, The Azuara impact structure (Spain):new insights from geophysical and geological investiga-tions, Geol. Rundsch. 81 (1992) 403–427.

[26] J.E. Warme, H.-C. Kuehner, Accretionary carbonate im-pact-ejecta spherules in the Upper Devonian Alamo Brec-cia, south-central, Nevada (abstr.), Geol. Soc. Am. Abstr.Progr. 29 (1997) A-80.

[27] J.M. Grajales N., D.J. Moran, P. Padilla, M.A. Sanchez,E. Cedillo, W. Alvarez, The Loma Tristes Breccia: a K=Timpact-related breccia from southern Mexico (abstr.), Geol.Soc. Am. Abstr. Progr. 28 (1996) A-183.

[28] J. Smit, Th.B. Roep, W. Alvarez, A. Montanari, Ph. Claeys,J.M. Grajalez-Nishimura, J. Bermudez, Coarse-grained,clastic sandstone complex at the K=T boundary aroundthe Gulf of Mexico: deposition by tsunami waves inducedby the Chicxulub impact? in: G. Ryder, D. Fastovsky, S.Gartner (Eds.), The Cretaceous–Tertiary Event and OtherCatastrophes in Earth History, Geol. Soc. Am. Spec. Pap.307 (1996) 151–182.

[29] B.J. Witzke, R.R. Anderson, Sedimentary-clast breccias ofthe Manson impact structure, in: C. Koeberl, R.R. Anderson(Eds.), The Manson Impact Structure, Geol. Soc. Am. Spec.Pap. 302 (1996) 115–144.

[30] G. Graup, Terrestrial chondrules, glass spherules and accre-tionary lapilli from the suevite, Ries crater, Germany, EarthPlanet. Sci. Lett. 45 (1981) 407–418.

[31] V.L. Sharpton, L.E. Marin, J.L. Carney, S. Lee, G. Ryder,B.C. Schuraytz, P. Sikora, P.D. Spudis, A model of theChicxulub impact basin based on evaluation of geophysi-cal data, well logs, and drill core samples, in: G. Ryder,D. Fastovsky, S. Gartner (Eds.), The Cretaceous–TertiaryEvent and Other Catastrophes in Earth History, Geol. Soc.Am. Spec. Pap. 307 (1996) 55–74.

[32] V.R. Oberbeck, The role of ballistic erosion and sedimenta-tion in lunar stratigraphy, Geophys. Space Phys. 13 (1975)337–362.

[33] W.C. Ward, G. Keller, W. Stinnesbeck, T. Adatte, Yucatansubsurface stratigraphy: implications and constraints for theChicxulub impact, Geology 23 (1995) 873–876.

[34] H.J. Melosh, Impact Cratering: A Geologic Process, OxfordUniversity Press, New York, 1989, 245 pp.

[35] P.H. Schultz, Atmospheric effects on ejecta emplacement,J. Geophys. Res. 97 (1992) 11623–11662.

[36] O.S. Barnouin-Jha, P.H. Schultz, Ejecta entrainment byimpact-generated ring vortices: theory and experiments, J.Geophys. Res. 101 (1996) 21099–21115.

[37] O.S. Barnouin-Jha, P.H. Schultz, Lobateness of impactejecta deposits from atmospheric interactions, J. Geophys.Res. 103 (1998) 25739–25756.

chicxulub impact ejecta from albion island,...

Documents