GeologicalSociety of America SpecialPaper 247 1990 Late Cretaceous-earlyEocene mass extinctions in the deepsoa Ellen Thomas Departmmtof Eanh and EnvironmentalSciences, Wesleyan University, Middletown Connecticut06457, and TlnmesScience Center,New London, Connecticut 06320 ABSTRACT Upper Maashichtian through lowermost Eocene deep-seabenthic foraminiferal records from Maud Rise (Weddell Sea,Antarctica) demonstratethat there was no mass extinction of theseorganisms at the end of the Cretaceous.Ihere is no significantdrop in diversity acrossthe Cretaceous/Tertiaryboundary, butjust abovethis boundary there is a peak in relative abundanceof speciesthat may indicate low-nutrient conditions, probably reflecting the decreasein food supply to the ocean floor resulting from the large-scaleextinction of surface-dwellingprimary producers. In contrast, there was a global extinction of bathyal to abyssalbenthic foraminifera at the end of the Paleocene, occurring in fewer than 25,000yr at Maud Rise.Many benthicforaminiferal speciesthat had originated during the Cretaceousbecame extinct, although there was no coeval mass extinction (of comparable importance) of terrestrial organismsand planktonic marine organisms.After this extinction the diversity of benthic faunas on Maud Rise was low for about 260,000yr, and during the period of low diversity, the assemblages were dominatedby speciesthat may indicate the presenceof abundantorganic carbon, and possibly low concentrationsof dissolvedoxygen. The dominancezuggests that the Paleocene/Eocenedeep-sea benthic foraminiferal mass extinction was caused by a decreasein oxygen content of the waters bathing the lower bathyal reachesof the sea floor. Such a changecould have beencaused by a changein the circulation patternsof deep waters: these waters would becomedepleted in dissolvedoxygen if there was a change from predominant formation of deep waters at high latitudes (cooling and sinking) to initiation of, or a strong increaseof, formation at low latitudes(evaporation and sinking). Thus, one of the largest Phanerozoicextinctions at the Earth's zurtaceis not reflectedby the deep-waterforaminifera, and the largest Cenozoicextinction event in the bathyal-abyssalrealm of the oceansis of little importanceto surfacebiota: even someof the largestextinction eventsthat we know do not reachall environmentsof the Earth. INTRODUCTION "buffer to extinction"(Sheehan and Hansen,1986; Arthur and others,1987), although some authors suggest that bunowing or- At the end of tle Cretaceous,landdwellers and surface- ganismsunderwent a massextinction (Wright lt Hsti and others, dwelling organismsin the oceansunderwent one of the largest 1984,p. 335).Deepwater ostracodes have been reported to have extinctionevents of the Phanerozoic(e.g., Thierstein, 1982; Cle- had a "faunal crash" at the end of the Cretaceous(Benson and mens,1982; Russell, 1982), but extinctionrates in deep-seaben- othen, 1984),but few data areavailable for faunasliving just thic organismsare not well known (Culver, 1987; Thientein, after that "crash,"so mass-mortalitypattems for that group have 1982;Hsii, 1986).Benthic organisms in the deepoceans have not been well established(Steineck, personal communication, beensaid to showlittle or no changeacross the Cretaceous,/Ter- 1989;Benson and otlers, 1985, Fig. l). tiary boundaryQlsii, 1986);detritus feeding has been seen as a Benthicforaminifera, which supplythe mostabundant fos- Thomas, E., 1990, l,ate Cretaceous-early Eocene mass extinctions in the deep se4 lz Sharpton, V. L., and Ward p. D., eds., Global catastrophesin Earth history; An interdisciplinary conference on impacts, volcanism, and mass mortality: ceological Societybf America Special yapr j.+7. 481 482 E. Thomas sils of deep-water organisms,were reportedly little affectedby the (65"9.629'5,Io12.296'E, present water depth2p14 m) were mass extinction at the end of the Cretaceous (Douglas and drilled on Leg ll3 (January-March1987) on Maud Rise,an Woodruff, l98l; Thientein,1982; Culver, 1987),but this obser- aseismicridge at the eastemend of the Weddell Sea (Barker vation is based on few quantitative data sets. Many earlier andothers, 1988; Fig. l). Site689 is on the northeasternside of worken on benthic foraminifera described the great similarity of the ridge near its crest,Site 690 is on the southwesternflank, Maastrichtian and Danian (lower Paleocene)faunas (e.g., Cush- I 16 km to the southwestof Site 689. LowermostMaastrichtian man, 1946). More recently, differences have been recognized, through Pleistocenebiogenic sediments were recoveredat both although estimatesof extinction rates vary widely (18 to 67 per- sites(Fig. 2).T\e UpperCretaceous through lower Eocenecon- cent species extinction; Beckmann, 1960; Webb, 1973; Beck- sists of calcareouschalks and oozes;fine-grained terrigenous mann and others, 1982; Dailey, 1983; Widmark and Malmgren, materialis presentin partsof the sectionat Site 690 (Barkerand 1988; Keller, 1988b).Several of the estimates(Beckmann, 1960; others,1988, p. 190-l9l). Webb, 1973; Beckmannand others, 1982) must be seenas max- Paleodepthsof the sitescould not be estimatedusing simple imum estimates,simply becausethey were derived from compari- thermal subsidencemodels because Maud Rise is an aseismic sons of faunal lists for the Maastrichtian and Danian. Therefore, ridge; benthic foraminiferalfaunas indicate latest Cretaceous- theseestimates include last appearancesthat occurred a consider- Paleocenedepths of 1,000to 1,500m for Site689, and 1,500to able time before the boundary. 2,000 m for Site 690 (Thomasin Barkerand others,1988; Environmental conditions at the end of the Paleocenecon- Thomas,1990). For severalsites drilled on ODP Leg 114(Sites trast with those at the end of the Cretaceous.At the end of the 698-702; seeFig. I for locations),paleodepths could be esti- Paleocene there were no extinctions among shallow-water and mated using simple thermal subsidencemodels; overall, these surface dwellen comparable in size to the extinctions at the end deptls showedgood agreementwith depthsderived from faunal of the Cretaceous: it was a period of below-average extinction data (Katz and Miller, 1990). Comparisonof the Maud Rise rates (Raup and Sepkoski, 1986). The diversity of marine plank- faunasand the Leg 114 faunasconfirms the depth estimateof tonic microorganisms such as dinoflagellates, calcareousnanno- 1,000to 2,000m for Sites689 and 690 duringthe lateMaastrich- plankton, and foraminifera was increasing after the middle tian-early Eocene.In this study the following bathymetricdivi- Paleocene(e.g., Oberhaensli and Hsii, 1986). At the sametime, sionswere recognized,in agreementwith Berggrenand Miller however, deep-seabenthic foraminifera underwent their largest (1989):neritic-<2OO m; upperbathyal-2O0 to 600m; middle known extinction of the Cenozoic (e.g., Beckmann, 1960; Braga bathyal-600 to 1,000m; lower bathyal-1,000 to 2,000 m; and others,1975;Schnitker,l9T9; Douglasand Woodruff, 1981; upper abyssal-2,000 to 3,0fi) m; lower abyssal) 3,000 m. 1-alsmaand Inhmann , 1984; Culver, 1987), and this extinction Therefore,both sitescan be placedin the lower bathyalinterval has remained unexplained. Some authors suggestthat primary for the tine periodstudied. productivity decreasedat the end of the Paleocene(Shackleton The biostratigraphicinformation was obtained from cal- and others, 1985), and this could have caused the extinction of careonsnannofossils (Pospichal and Wise, 1990a,b, c); only a deepwater dwellen; the suggestionhas not been widely accepted, few datum levelsof planktonic foraminiferawere reliable age however (Miller and others, 1987b; Katz and Miller, 1989). indicaton at thesehigh latitudes (Stott and Kennett, 1990a; One purpose of this study was to contribute to the knowl- Huber, 1990). In addition, data on the paleomagneticrecord *fabriC'ofthe edgeofthe extinction at the end ofthe Cretaceous wereprovided by Hamilton (1990) for the Mesozoic,and Spiess by collecting a quantitative data set on ranges and abundance (1990) for the Cenozoic.The stratigraphicinformation supplied patterns of deep-seabenthic foraminifera from Maud Rise (Wed- by all theseauthors is compiledin Figure2. dell Sea, Antarctica). Another purpose was to compare the In this chapter,dzta ue presentedfor the interval between changesin faunal composition of deep-seabenthic foraminifera at 140 and 260 meten below seafloor (mbsf) at Site 690 and 200 the CretaceouslTertiary (K/T) boundary (a time of collapse of to 260 mbsf at Site 689, correspondingto upper Maastrichtian the primary productivity; Arthur and others, 1987) with faunal (Nephrolithusfrequens Zone; Pospichal and Wise, 1990a) changesat the end ofthe Paleoceneat the samelocation, to asses throughlowermost Eocene (CP9; Pospichal and Wise, 1990c). whether the pattems of faunal changewere similar. This informa- Absoluteages were derivedfrom crossconelation of the paleo- tion should be important in evaluating whether mass extinctions magneticand calcareousnannofossil data with the geomagnetic reach all environments from the surface to the lower bathyal polaritytime scaleof Berggrenand others(1985). At Site689 areasof the ooeans,or whether thesetwo environments (and thus there are unconformitiesat the K/T boundary (Zone CPla, their inhabitants) are largely decoupled. severalhundred thousandsof yean), in the middle Paleocene (Zones CP4 through CP5), and in the topmost Paleocene MATERIAL AI\D METHODS throughlower Eocene(at leastZones CP9 throughCPI I about 6.6m.y.; Fig. 2). Thereis no Ir anomalyat