Oct 11, 2007 ... (damonds) overlyng the carbonate results (th~n l~nes). Occas~onal samples were analyzed n dupi- cate, nclud~ng light po~nts between 35 ...
The Little Ice Age and Medieval Warm Period in the Sargasso Sea Lloyd D. Keigwin Science, New Series, Vol. 274, No. 5292. (Nov. 29, 1996), pp. 1504-1508. Stable URL: http://links.jstor.org/sici?sici=0036-8075%2819961129%293%3A274%3A5292%3C1504%3ATLIAAM%3E2.0.CO%3B2-9 Science is currently published by American Association for the Advancement of Science.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/aaas.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact [email protected]
http://www.jstor.org Thu Oct 11 12:25:49 2007
deposition rates at this site because the
The Little Ice Age and Medieval Warm Period in the S ~ K ~ ~Sea S S O
signal has not heen diluted o n the seafloor by ~ ~ p w a rmixing d of older foraminifera. Additional A M S ''C measurements at each . .subcore of BC-004 . . -have. -been , - used to Lloyd D. Keigwin develop a n age model (big. 1C) tor stackSea surface temperature (SST), salinity, and flux of terrigenous material oscillated on lng tlie data (Fig. IF) (8).Evidently plston millennia1 time scales in the Pleistocene North Atlantic, but there are few records of core GPC-5 failed to recover tlie upperHolocene variability. Because of high rates of sediment accumulation, Holocene oscil- most -15 cm of the seafloor at this lations are well documented in the northern Sargasso Sea. Results from a radiocarbon- location. dated box core show that SST was -1 "Ccooler than today -400 years ago (the Little Percent C a C O , near the top of the box Ice Age) and 1700 years ago, and -1 "Cwarmer than today 1000 years ago (the Medieval core decreases fro111 maximum values beWarm Period). Thus, at least some of the warming since the Little Ice Age appears to be tween 1000 and 500 years ago to the art of a natural oscillation. most vronounced mlnimurn of the nast 10,000 years, centered o n -400 years ago ( F ~ g s .1 F and 2 ) . This minimum and the preceding nlaxlmunl are e ~ ~ u i v a l e in n t age With the exceptLon of nearshore anoxic ha- Piston core KNR31 G P C 5 lins a r e s e r ~ ~ o i r - to tlie cliniate events loosely knoa.11 as the sins ( 1 ), climate data from marine sediments corrected accelerator mars spectrometer Little Ice Age (LIA) and the Medieval generally lack sufficient time resolution to be (AictS) ItC age of 860 years at the top ( 6 ) . Warm Perlod (ivIWP), respecti\~ely ( 9 ) . increase in carconit7ared dlrectlv to instrumental ohserva- In contrast. each of two sub cores i A and Topether with the slieht u tions. Thus, there hns been n gap in our B) of BC-004 has zero or negative plank- bonate in the upper few centimeters of the kno\vledge of tlie ocean-climate system he- tonic fi>ra~niniferal ages after reservoir cor- core, the minimum and maximum define tween tlie m~llennial-scaleclimate clianges section (Fig. 1, h and B, and Tahle 1 ) . the tlilrd of three cycles during the last resolved in the best sediment cores trom the This age difference must result from better 5000 years o n the Bernluda Rise. This open ocean anci decadal scale clinnges 011recovery of the sedinlent surface by BC- broad interval is often referred to as the served in instrumental data. This cap occurs 004. Zero or neeative aces reflect excess ~7eriodof Neoelaciation ( I @ ) , ~vhiclifolon tlie century time 'scale, tlie very time scale lfC produced 1iy atmospheric nuclear lo~veda n early Holocene period of Lvarmer on a h i c h anthropogen~cwarming is thought weapons testing and attest to the high cl~rnate(tlie Hypsithermal; Fig. 2). It is to be occurring (2). Hence, it 1s i~uportantto document natural climate variahility 11-1 orTable 1. Results of accelerator Inass spectrometer radiocarbon datng of Bermuda Rlse box core der to effects of al,tliropo. HU89-038-BC4 at the Natonal Ocean Sclences Accelerator Mass Spectrometer Faclty, Woods Hole. forcillg. Here I report tlie record of Massachusetts. All ages In years before present. climate change for the past fe\v ~nillennia from the Bermuda Rlse in the northern SarCalendar gasso Sea, a location ahere century-scale Depth No, Fracton ReSenrO1r Cale~idar age range, Measured corrected fcm) modern resolution is poasihle. age aae age? 2 I(J " T h e Bermuda Rise is a remarkable archive of proxy climate ~nformation.R ates of secllme~itaccumulation as high as 200 cnl per thousand years are maintained by deep recirculating gyres ( 3 ) \vhicli focus Jetrital silt and clay at this locale ( 4 ) . Cycles of varlous sedimentary properties occurred in tlie late Pleistocene with a quasi period of 4000 years (5),extending through the latest deglac~ationanci into the Holocene (6). In the late Pleistocene these variations are found in calcium carhonate content, in stable Isotope ratlos of oxveen and carbon, and in CdICa ratios. , all of which are thought to reflect, in part, 0.5 5178 0.9513 400 2 40 0 0 0 clinnges in S S T and changes in the prc5321 0.9665 275 % 30 0 0 0
ciuction of North Atlantic Deep Water. Sampling of the last -10,000 years ( t h e Holocene) o n the Bermuda Rise had not heen sufficient to deternline vreciselv the state of the modern climate system wit11 respect to these millennial-scale cycles. T o address this question, I studied a box core (HU89038 BC-004) taken fro111 the same locat1011 as previous cores (7). l:'ioods HoleOceanograph~cnsttution bh200dsKole, Iv'A 02543 USA
2.5 3,5 4.5 5.5 6.5 7.5 8.5 10.5 13,5 45,j 48.5 50.5
5176 5177 5322 5172 5319 5171 5323 5324 5320 5325 5326 5327
1.0540 0.9855 0.9587 0.9080 0.8982 0.8963 0.9003 0.9034 0.8585 0.6921 0.6633 0.6669
1 1 5 % 30 340 t 30 775 2 40 860 t 40 880 -t 30 845 t 20 815 % 40 1230 t 20 2960 % 30 3300 t 30 3250 t 40
0 0 0 375 460 480 445 415 830 2560 2900 2850
'Correctecl by 4 0 0 years for t i e age of the surface ocean resewor accord~ngto (26).
\.OL. 2 74
29 NOVEh'lBER 1996
0 0 0 424 483 494 474 455 749 2731 3129 3059
0 0 0 385-452 461-503 475-510 457-490 428-478 723-777 2714-2747 3077-31 77 2984-31 15
-:Calendar ages have beet? calbrated
likely t h a t carbonate flux \\.as relatively constant a n d t h e flux of terrigenous sedirnent iielivereii to t h e Berins~iia Rise by deep currents increased during t h e L1.4, just as it iiiid during earlier carbonate minilna in t h e Holocene a n d durlng glaciation ( 4 ) . For b o t h t h e L I A and t h e carbonate minimum of -1300 years ago, increased flux of terrigenous clay a n d silt particles probably accounts for t h e significant increases in sedilnentation rates indicated hy t h e .4h/lS dates (Fig. 1 C ) . T h i s terrigenous sediment most llkely \\.as resuspendeci from tlie Scotian Rise during abyssal storms ( 1 1 ), or eroded tram t h e northeast scarp of t h e Bermuda Rise ( 1 2 ) . A t m o spheric stor~niness could force tlie Gulf Stream to be Inore energetic, in turn increasing t h e kinetic energy of t h e deep reclrculating gyres to \vhich it is linkecl (1 3 ) . Historical evidence ( 14 ) and geological evidence a t nearshore locations ( 1 5 ) point t o increased s t o r ~ n i n e s sduring t h e LI.4 in t h e N o r t h A t l a n t i c region. W h e t h e r t h e source of t h e terrigenoi~s dilutant was local (Bermuda Rise scarp) or more distal (Scotian Rise), tlie prelxiling deep flow w o i ~ l dtransport this sediment t o t h e plateau of t h e Bermuda Rise ( 3 ) . A t present, t h e geochemical evidence for is a1nhig~1chaliges i n SXL)W ous, so it is n o t k n o \ ~ nif tliere was actr~ally a c h a n g e in t h e source of d e e p waters associated \vith t h e LIX, as there n a s for s ~ r n i l a revents i n t h e Pleistocene (5, 16). I n general, it is t h o ~ ~ g thhta t any Holocene changes i n deep o c e a n n.ater inasses were small compared to those during iieglacial a n d older tllnes ( 1 7 ) . O s y g e n isotope ratios (8'") are a more direct proxy for cliinate c h a n g e iiuring t h e late Holocene in t h e Sargasso Sea t h a n percent carhonate. For isotopic analysis I chose t h e planktonic fi~rarninifera Globige~inoidesr~iber( w h i t e variety, 150 to 250 p i n ) . T h e white variety of this specles lives year-round i n tlie upper 25 in of t h e n o r t h e r n Sargasso Sea a n d has a relatively constant aniirlal mass f l u s a n d shell flux ( 1 8 ) . T h u s , of all planktonic foraminifera a t this location thls specles is most appropriate for reconstructing a n n u a l average S S T s 118). , , Oxygen isotopic ~ a r i a t ~ o nofs G . niher 111 BC-334 should reuresent o11ly t h e influence of SST a n d salinity variations i n t h e Sareasso Sea. Results a t e a c h sul3core generally 1-ary 111 concert \\.it11 tlie percent C a C O , . M a s ~ m u mS1"O values are associated nit11 m i n i m ~ ~percent m C a C O , (Fig. 1, D a n d E) AS they are earller 111 t h e Holocene (Fig. 2 ) . Increased S1'O and decreased percent C a C O , are consistent u . ~ t hclimatic cooling i n t h e Pleistocene o n t h e Berinrlda Rise (5) a n d a t o t h e r
Nortli A t l a n t i c locations. A t times, t h e S 1 ' 0 ressults s h o n e\.en higher frequency l-ariability t h a n t h a t of t h e carbonate record, especially during t h e uppermost maximum in percent C a C O , (Fig. 1, D and E). A l o n g with e v ~ d e n c eof slightly different phasing b e t ~ v e e n t h e carbonate and 6 l S O records in t h e Pleistocene ( 5 ) , unlinknig of t h e two by just -1000 years ago suggests t h a t t h e 6 1 S 0 results are n o t a n artifact of carbonate dissolution. T h e occurrence of identical oscillations of
6"O \ d u e s i n G . ruher a n d t h e solutionresistant Glohorotalia inflatic a t this location during t h e prolonged deglaclal carbonate minimum provides additional e\+ dence t h a t G . ruber is recording a primary surface water signal ( 6 ) . T h e 42-year series of hydrographic d a t a a t Bermuda S t a t i o n "S", which is -700 k m t o t h e southn-est of BC-004, provides a n important baseline for interpreting t h e geological record because within this broad region hydrographic properties s h o n
I 0 40-
BC-004D . , , 20 30 40
~ 10 . 60
Fig. 1. Sedimentological. oxygen isotopic, and radiocarbon data on Bermuda Rise core HU89038 BC-004. (A and B) Weght percent carbonate in two subcores of BC-004 wth results of A M S "'C dat~ngon mixed planktonic foraminfera (see Table 1 ) . The radiocarbon dates have been corrected by 4 0 0 years to account for the age of surface waters n which the foraminifera grew. Dates near the core top w~thzero or negatve ages after reservor correcton are not shown. One '4C analys~sn each subcore gave a fraction of modern carbon > I . ~ndlcatngthat the negative ages reflect the presence of bomb '"C. (C)Age-depth plots for the two subcores show~ngthe age models adopted for stacking the data. Note that the hghest rates of sed~mentatonoccur where percent carbonate IS a m n m u m . supporing the nterpretation that d u t o n by clay and silt particles drives the percent carbonate results. (D and E) Oxygen soto ope results on the surface-dwelng planktonc foram~nferaG. ruber (damonds) overlyng the carbonate results (th~nl~nes). Occas~onalsamples were analyzed n dupicate, ~nclud~ng light po~ntsbetween 35 and 50 cm in (D).(F) Percent carbonate results the ~sotop~cally on a calendar-year age model based on the control ponts n (C).Open symbols. BC-004A: solid symbols. BC-004D.
little geographic variability ( 1 9 , 20). Interdecadal variability in monthly normalized property anolllalies are ell kno\vn at Station "S" 119-22)., , hut even o n a n annual average hasis a major event like the late 1960s decrease in S S T and increase in salinity stands out (Fig. 3 ) ( 2 3 ) . T h a t climatic extreme and similar enisodes ;it other times are associateii ~ i t lninilna h in the North Atlantic Osc~llation,strollg westerlies at 1on.e~illidille latituiles ( 2 4 ) , and s o i ~ t h w e s t ~ a rinovement d of storm centers in the North Atlantic (25). In order to gauge the infli~ellceof the annual varial3jlity of S S T and salinity o n the oxygen i'sotope ratios that might be
fi>raminiferal isotopic data inostly in terms recorded by surface-dwelling fi>raminfera at Station "S", I c a l c ~ ~ l a t e i l of S S T change. W h e n the S1"O data from each sul3core the 6 % value of calcite precipitateil in are plotted together o n a calendar time oxygen lsotopic eiluilihrium ~ i t seawater h (Fig. 3 ) ( 2 6 ) . As expected, the effects of scale ( 2 7 ) , it is clear that the same features decreased S S T anil lllcreaseil salinitv ilur- are present in each subcore ( 2 8 ) . Within ing the late 1969s cornblned to increase each subcore S I Q values reach a minithe 6 ' 9 value of eouilibrium calcite. mum -590, 900, and 1100 years ago (Fig. Over the full 42-year series, linear regres- 4 A ) . Using these data, I solved the paleosions b e t ~ e e nthe SST, salinitv, and S1'O telllperatllre equation ( 2 6 ) after applying values ahon. that temperature accounts for Deuser's disecluilibri~~rn correction (18 ) of about t ~ o - t h i r d sof the isotonic signal ( r L +0.2 per mil to the 6 % value of G . ruber = 9.61), whereas salinity accounts for oneand assuming that the average salinity was third (? = 0.29). Thus, by comparison to 36.5 per mil. I then stacked the temperasea surface changes during the past several ture proxy data from the tn.o subcores by decades, it is reasonable to interpret the averaging results in 50-year bins (Fig. 4B). In general, these results indicate that there have been century-scale changes in Fig. 2. Holocene climate proxy S S T of 1" to 2°C throughout the past fen. data from the Bermuda Rise, norththousand years in the Sargasso Sea. In the ern Sargasso Sea. (Top) Oxygen last half of the record there u.as a l . i ° C isotope rato of the surface-dwelling oscillation frolll a lllinillluln S S T 1599 to planktonic foram~n~feraG. ~ i b e i 1790 years ago to a maximum 900 to 1009 from BC-004 and nearby gravity years ago, to a nlininluln 390 t~ 490 years core EN1 20 GGC-I . The box core ago. Since the Little Ice Age, SSTs in the data are a stack of the data In Fig. northern Sargasso Sea increased by -1°C. 4A at 1 -cm spaclng, and the gravty Actual S S T changes may ha\-e been even core data are at 2-cm spaclng. Where the two cores overlap. the greater t h a n ~ n d i c a t e din Fig. 4B, in that EN120 GGC-I ....... ........................................................................................................ GGC-I data were added to the box the sediment may have been ~llixeddiffercore stack. Because of the different entially by h u r r o ~ \ ~ i n(-5 g cm) as sedisampllng intenias between the two mentation rates changed, and because cores and the Increasing rate of stacking the S l h O data may have attenusedimentation since the early Holoated the signal. From the raw ?i1" data of cene lninlmum (G), data for the last BC-004D (Fig. 4 A ) , calculated S S T 350 3000 years have higher resolution. years ago mas 21.i°C, about l . i ° C colder (Bottom) Percent carbonate re0 ** than the rnodern annual average. sults from core GPC-5 (solid cirDepression of SSTs by l o to 2°C during cles), GGC-I (open circles), and 20 ............................................................. Hypsithermal BC-004 (triangles).Note that the the LIA is consistent with other proxy 1% MWP three major carbonate minima data fro111 the Sargasso Sea region and is (-3500, -1500, and -400 years probably part of a much larger climatic ago)are closely matched by oxygen pattern ( 2 9 ) . T h e 800-year record of coral sotopc maxima (toppanel)b~ltthat Neoglaciation ............................................................ growth near Bermuda has been interpreted the isotopic data also contain a 0 1 ~ ~ , i ~ , , I I . ~, , ~ , I ~ ~ in terms of S' S T cooling of this magnitude, higher frequency signal. 0 2000 4000 6000 8000 10000 \vhich was ind~lcedat least in part by Calendar years before present \vind-driven vertical mixing and heat flux changes (30). Far to the west in the Florida Strait, Ai4C and 6'" values o n coral Fig. 3. Annual average also indicated that SSTs were lower by l o 36 25 -1 00 00 data from Station S 00 Sallnlty oo to 2°C between -A.D. 1680 and 1750 0, oo~O near Bermuda (3236 50O O"oooOOOo OooOo during the LIA ( 31). ooO 0 0000 oo~OOo 10'N 62'30'W) The Abrupt, century-scale changes in SSTs - -0 75 temperature and sallnlty 36 75 --24'C in the Sargasso Sea may have ~nfluenced data are scaled ~nproae 0 0 g 0 0 climate do\vnstreain in \videspread regions rn 0 SST porion to ther effect on ; 37 00.- to the east. It is t h o ~ ~ g that 0 0 0 n s P o # o h t S S T changes -0 50 2 the 6 50 of calclte pre -23C 0 o ~ O n 0 0 qp.n*nbo.. 0 in the North Atlantic (as well as global clpltated In equbr~um 37 25 l 0 l 0 . l w~th seawater (solld '" changes) ~ n d u c echanges in African rain-0 &I80 data) to show that the 37 5 0 - - 2 2 0 ~ 0 . fall ( 3 2 ) , and the lake level hlstory of 0 2 5 l l temperature effect on tropical African Lake B o s ~ ~ m t m~ndicates i * 6 5O I S about twce that 37 75 that extended drought and lowest lake of the s a n t y effect over levels in the Holocene occurred during 0 00 I I I I I recent decades at th~s 1950 1960 1970 1980 1990 2000 the LIA ( 3 3 ) . Summer tenlperature estilocatlon Note that unke mates d e r i ~ ~ efro111 d tree rings in northern lnost locat~onsthe ternYear perature and s a n t y from thls part of the Sargasso Sea are only slightly correlated (or not at all) Fennoscandia shoal that the MWP conHowever durng a pronounced cllnate event ~ k the e mlnltnuln In the North Atantc Oscllatlon of the late sisted of two events, the first in the 10th and 11t h centi~rlesand the second 111 the 1960s the SST and s a n t y combne to produce a robust Increase In 6150 values
. : . . . .* ........ .. .
early 15th century ( 3 4 ) . Furthermore, these are bracketed by cold events ( 3 4 ) that seem to correlate mith glacier exvansion in southern Norway (35). During the LIA, elaciers in southern Norwav reached their ireatest extent of the past 9b00 years (35). Thus, as is sunlrnarised in Fig. 4 A , it appears that the high-resolution S S T record from the Bermuda Rise is consistent mith the lu~lltidecadaltrends in the highest resolution records of cllmate proxy data o n land to the east. Over the course of three millennia, the raqge of S S T variability in the Sargasso Sea is o n t h q order of twice that measured over recent 'decades (Fig. 4B). Increased l~arianceof climatic spectra o n longer time scales has been demonstrated for the climate system in general ( 3 6 ) , and for the North Atlantic region in particular (37). Although forcing for climate change o n millennia1 and centennial time scales is still poorly understood, the North Atlantic Oscillation I N A O ) inav, t~rovidea use. ful model for interpreting long-term S S T changes in the western Sargasso Sea. If N A O minimum conditions mere more persistent during the LIA they could account for m a n r of the Bermuda Rise observations and the coral observat~ons,including S S T depression, salinity increase, increased pumping of nutrients to the sea surface (increasing coral growth rates), and increased terrigenous load in deep r e c i r c ~ ~ l a t i ngyres g as a result of the south-
westward shift in storm tracks. Dickson et
al. (22) c o n c l ~ ~ d ethat d o n interdecadal I N A O ) tiine scales convection in the Sargasso Sea jforluation of subtropical mode water) is in phase mith convection in the Greenland Sea, and out of phase mith conr~ectionin the Labrador Sea. T h i ? Dattern is similar to the sort of conrrective dipole that has been proposed to operate o n glacial-interglacial time scales ( 3 8 ) in order to account for the change from deep to intermediate depth ventilation In the North Atlantlc (17, 39). It \vould he interestine to know if the same strle of climate and ocean variability occurs In the Korth Atlantic across the fi111 spectrum of 10' to 10' years. Because climate events like the LIA and MWP were of long enough duration (decades to centuries) to be resolved in Bermuda Rise sediments, and because the chanees described here for surface waters over the Berlnuda Rise are probably typical of a large part of the \vestern Sargasso Sea, they most likely reflect climate change o n the basin or hemispheric scale. Regardless of the exact cause for the LIA, the MWP, and earlier oscillations, the warming during the 20th century (O.S°C) ( 2 ) is not unprecedented. However, it is important to distinguish natural cllrnate change from anthropogenic effects because human influence may be occurring at a time when the climate system is o n the warming limb of a natural cycle. D
Fig. 4. (A) Oxygen lsotope ratlos of the surface dweling planktonic foramin~fera G. ruber from Berlnuda Rise BC-004 plotted ~ e r sus calendar age. Open symbols, BC-004A; solid symbols, BC-004D. Bars Y (5 above the abcissa are a glacation In S. Norway -= -r cold Fennoscandian summers schematic representaton 0'00rn A rru r. warm Fennoscandian summers of proxy data for episodes I " " I " " I " " I " " I " " of glacial expansion In southern Norway (35)and summer temperature varlability in Fennoscanda reconstructedfrom tree rings (34).For the tree ring data, temperature m a m a and minima are shown by thicker bars. For the glacier data. the bars represent 1 I~ 2 1 1 ~ ~ two early perods of expan0 500 1000 1500 2000 2500 3000 sion (but not to LlA mits), Calendar years before present followed by the range of age estimates for attainment of the LIA maximum (35).These terrestrial data, which are downstream of the Norih Atlantic, are generally consistent wlth the Z i 8 0 data (maxmacorrespond to cooling, minma to warming). (6)Sea surface temperatures calculated from the 8 ' 9 data in (A), after averaging the data in 50-year Intervals, plotted w~ththe annual average of SST measured at Station "S" snce 1954 (from Fig. 3).Although, as dscussed In the text, about one-thrd of SST varablity calculated from 8'" values (beforestacking)may actually reflect salnity change in the SargassoSea, t IS clear that on centenna and millennia1tme scales,SST \/arabtyhas been greater than has been measured over the past four decades at Station "S."
29 NO\'EMBER 1996
REFERENCES AND NOTES 1. N. G. P~sias,Quat. Res. 10, 66 (1978); A. JuilletLeclerc and H. Schrade:. Nature 329, 146 (1987); J. P. Kennett and B. L. lngram ibid. 377 510 (1995);K. A. Hughen. J. T. Overpeck, L. C. P e t e r s o n 5 Trumbore, ibid. 380, 51 (1996). 2 P. D Jones. T. M L Wigey. P. B. Wr~ght.Nature 322,430 (1986); B D. Santer et al. , Clim. Dynam. 12, 77 11995). B. D Santer et a/., Nature 382, 39 11996). 3 . E. P. Lane and C. D. Hollster. Mar Geol 39, 277 (1981); W. J McCartney and M. S McCartney. Rev. Geophys. 31, 29 (1993). 4. M. P. Bacon and J. N. Rosholt Geochim Cosmochim Acta 46, 651 (1982) D 0 . Suman and M. P Bacon. Deeu-Sea Res. 36. 869 11989) 5 L. D. ~ e g w ; nand G. A. Jones, J. Geophys. Res. 99, 12397 (1994). 6. , Deep-Sea Res. 3 6 845 (1989). 7. BC-004 was recovered from 33"1.6'N, 57'36.7'W by CSS Hudson at 4418 m on the undulatitig plateau of the northeast Bermuda Rise, within 1000 m of KNR31 GPC-5 (6) and EN120 GGCl (17j.The box core was subcored uslng p a s t c plpe w t h adameter of 11 4 c m 8. As BC-004A 1s generally better dated than BC-004D, the two were correlated and ages assgned to the latter to bring the carbonate curves n t o llne (Flg. 1 F). For the last 500 years, BC-004D IS more ntensely dated, and an age model was chosen for the interval from 5.5 to 10.5 cm to be w~thlnthe 1u age errors (Table 1).Although the age of the youngest carbonate mlnmum dlffers sl~ghtlybetween the two subcores, the age model br~ngs61P0 results into good agreement (Fig. 4A). 9 In t h ~ sreport these names are used ~nformallyto denote cool~ngseveral hundred years ago, and warmlng about 1000 years ago. Although there IS general agreement that there was coolng and glacal expansion n many reglons dur~ngthe LIA Interval, and warmng durng the MWP Interval, t IS by no iiieans certa~nthat spec~ficevents or even these intervals were synchronous and worldw~de.Contrast, for example, M K. Hughesand H. F Diaz Clim. Change 26, 109 (1994) w ~ t hJ. M. Grove and R. Switsur, ibid. 26, 143 11994).See also R. S. Bradley and P. D. Jones, Cbmate Since A.D. 1500 (Routledge. London, 1995). 10. S C. Porter and G. H. Denton, Am. J. Sci 2 6 5 1 77 (1967). 11. Br~efevents of Intense deep current actv~tyand sedlment resuspenson known as abyssal storms have been observed off Nova Scota by C. D Holster and I. N. McCave [Nature 309, 220 (1994)l. 12 E. P, La~ne,W!. D. Gardner, M. J R~chardson M Kominz Mar Geol 119 159 (1994). 13 L V Worthngton On the Noith Atlantic Circulation (Johns Hopkins Oceanogr Stud. 6, Johns Hopkns University, Balt~more 1976). 14 H H. Lamb, Quat Res. 11 1 (1979) 15. H. C. Hass, Mar. Geol 111 361 (1993) D J. W Piper M Feetham, J. Goulden, paper presented at the Fifth International Conference on Paleoceanography, Halifax N0.a Scota 10 to 14 October 1995, p 58. 16 E. A Boyle and P D. Rosener,Palaeogeogr Palaeoclimatol Palaececol. 89 113 (1990), D. W. Oppo and S. J Lehman, Paleoceanography 1 0 901 (19953. 17. E.A. Boyle and L D. K e g v ~ nNature 330 35 (1987). 18 W!, G. Deuser, J. Foraminifera1Res 17 14 (1987) 19 L D. Talley Ptiysca D, In press. 20 T. M Joyce and P. Robblns, J Clim in press. . Talley and ~M E. Raymer, ~ J. Mar. Res~ 40 21. L. D. (suppl.3,757 (1982). VV J Jenkins ibid., p 265 22. R. D~ckson,J Lazier J. Meincke P. Rhnes, Prog. Oceanogr. In press. 23. T h s part of the Sargasso Sea IS u n k e most other locations in that surface temperature and salln~tyare not sgnif~cantlycorrelated (19, 20). Thfs probably results from latent heat loss associated bvth storminess [(19), see also (24)] Joyce and Robbns (20) also pont out the contributon of summer r a n s ~ , ~ ~ h l c h begn near Bermuda after establishment of the seasonal thermocl~neWith regards to the SqaO\!slues
24 25. 26
the nfluences of temperature and s a n t y changes in the Sargasso Sea surface are addltl!!e, whereas they tend to cancel each other at other ocatons. D. R. Cayan, J. Clim. 5. 354 (1992); Y Kushnr, ibid. 7, 141 (1994). R. R D~cksonatid J. Nam~as,M oil, lI/eather Reg. 104, 1255 (1976) E q ~ ~ b r i ucalcte m precpitaton was calc~~lated uslng annual average Stat~on S" data, the North Atlant~c salin~ty-SiaOrelat~onsh~p [H Craig and L I Gordon. Sympos~uniot?Marme Geoclieii-r~stiy(Occas. Pub1 3, Narragansett Marine Laboratory, Naragansett. Rl, 1965i1,and Shackleton's paleotemperature equaton [N. J. Shackleton. CNRS Colloq. 219, 203 (1974)l. Usng the calbrat~onof M Stuver and P J. Re~mer, Radocat-bon 35, 215 (1993). Subcores A and D $,wereanalyzed on tLvo different mass spectrometers (a VG Pr~smand a partally ailtomated VG Micromass 903. respectvely).Altiiough there mght be a slight ntercal~brat~on d~fferencebe-
31. 32. 33. 34.
tween the two nstruments, the small differences among the S'aO results most lhkely reflect sedmentap] processes such as b~oturbat~on. All data are arch\!ed at the N O W N G D C World Data Center-A for Paleoclimatology ([email protected]
go!!). R S. Bradley and P. D Jones, Holocene 3, 367 (1993). J. Patzold and G. Wefer, paper presented at the Four?il International Conference on Paleoceanography, Kiel, p. 224 (1992):R. B. Dunbar and J E. Cole. Coral Recot-;is of Ocean-Atmosphere Variability (Unversty Corporat~onfor Atmospheric Research, Boulder. CO, 1993). E.M. Druffel, Sc~ence218, 13 (1982). C. K. Foland, T N. Palmer, D. E. Parker, Nature 320. 602 (1986) M R. Tabot and G. Delibr~as,ibid. 268, 722 (1977); Earih Planet. S o Lett 47, 336 (1980). K. R Br~ffaet a / . Nature 346, 434 (1990); Clim Dynam. 7. 111 (1992)
A Combined Experimental and Theoretical Study on the Formation of Interstellar C,H Isomers
35. J. A Matthews, Holocene 1 , 219 (1991). 36. J M M~tchellJ r., Ouat. Res 6. 481 (19761 37. J. E.Kutzbach and R. A Bryson, J. Atmos. SCI. 31 1958 (1974). 38. S. J. Lehman and L D K e ~ g w ~Nati,re n, 358. 198 (1992). 39. D. W Oppo and S J. Lehman. Science 259. 1148 (1993) 40, I thank C E Franks and E Roosen for teciincal assistance, R. Bradley, E Druffel, and R Dunbar for cornments on the manuscript, M. McCartney, T. Joyce, W. Schmtz, E. Boye, S Lehman, and W. Jenk~nsfor d~scuss~ons, and I. Hardy K Moran, and the Bedford n s t t u t e of Oceanography for help In acquir~ngand archlvng BC-004. T h s ,work ,was funded by the NOAA Atlant~c Cl~mateChange Program. 21 August 1996; accepted 21 October 1996
molecule, acetylene, to synthesize hyclrocarboll radicals via a single atom-neutral collision in interstellar en\~ironments.T h e circumstellar shell of IRC+10216, for example, contai~lsCIHz as \yell as C('P,) reservolrs a t distances of 1014 t o 10" m from R. I. Kaiser, C. Ochsenfeld, M. Head-Gordon, Y. T. Lee, t h e central star (j), a n d forniation of C,H A. G. Suits via reaction 1 is feasible. O u r investigations also provide dynamical information The reaction of ground-state carbon atoms with acetylene was studied under single- o n t h e elementary steps to C'H isomers. collision conditions in crossed beam experiments to investigate the chemical dynamics T h e laboratory data strongly ciepend o n of forming cyclic and linear C3H isomers (c-C,H and I-C3H, respectively) in interstellar t h e structures of tlie initially formed C3H, environments via an atom-neutral reaction. Combined state-of-the-art ab initio calcu- collision complexes, a n d therefore we first lations and experimental identification of the carbon-hydrogen exchange channel to both calculated t h e ah i ~ i i t i ogeollietries of e n isomers classify this reaction as an important alternative to ion-molecule encounters to ergetically accessible C,H2 isomers. W e synthesize C3H radicals in the interstellar medium. These findings strongly correlate with t h e n colnpareil our crossed-beam data a n d astronomical observations and explain a higher [c-C3H]I[I-C3H] ratio in the dark cloud experimental dy~lalliicswith those arising from distinct C3H, adducts. O n c e t h e isoTMC-I than in the carbon star IRC+10216. mers were identified, n-e determined t h e exit cliannels from C,H2 follolving a carbon-liyiirogen bond rupture t o c-C,H, or F o r more tllan two ciecades, networks of T M C - 1 a n d t h e circ~~llistellare n \ ~ e l o p e l-C,H, or both. Ab initio electronic structure calcularadiative association, clissociative recom- surrounding t h e carbon star I R C + 1 0 2 1 6 bination, a n d exothermic ion-molecule re- to impro\,e t h e fit to astrono~liicals ~ ~ r \ ~ e yt si o ~ l swere performed a t a level of theory actions have been postulated t o account ( 4 ) . These models, holvever, suffer from high e n o i ~ g hto predict relative energies of for chemistry in t h e interstellar medium sparse laboratory data o n reaction prod- all local minima a n d reaction exother(ISM) ( 1 ) . S L I C reactions ~ involve ubiclui- ucts a n d c a n n o t elucidate t h e contribu- micities to a precision of ahout 1 to 3 kJ tous radicals such as linear and cyclic C,H tion to distinct structural isomers such as mol-' ( 6 ) . T h e ilisc~lssionis limited to t h e (l-C,H, propynylidyne, and c-C3H, cyclo- l/c-C,H. Therefore, e v e n this refined n e t - triplet potential energy surface (PES) because no triplet C,H, minimum fulfills t h e propynylidene) ( 2 ) ; for example, addition a.ork does n o t explain tlie interstellar of Cf to C,H, yielding l/c-C,HH is c - C , H to l-C,H ratio of ~111ityi n cold requirements for intersystem crossi~ig(7). tliollght to Ah follo~vedby a subsecluent molecular clouds compared to 0.2 2 0.1 O u r ah initio calculations show t h a t raciiative association of l/c-C,Hf a n d HI around I R C + 10216. H e n c e t h e formatioli propargylene, HCCCH, is t h e global minirnu~llo n t h e triplet C,H2 PES a n d is t o c - C , H , - , a n d a final dissociative elec- of i~iterstellarC,H isomers remains to be l ~ o u n dby 385.4 kJ mol-1 ~ v i t l irespect to tron-ion reconibination forming l/c-C,H resolved. a n d two liyilrogen atonis or H;. T h i s In this report, we present combined t h e reactants (Fig. 1 ancl Table 1 ) . T h e framen-ork, liolve\.er, c a n n o t reproduce high-le\,el ab initlo calculations ancl structure lias a n almost linear C-C-C a n observed number densities anii isomer ra- crossed-beam experiments o n the atom- gle of 171.9" and a torsion angle hetn-een tios. Fueled by recent kinetic studies of neutral reaction 1 to interstellar C,H iso- t h e two hyilrogen bonds of 88.0". Its C2 .: neutral-neutral reactlolls mers via C3H2interme~liates: b a~lierless,fast symmetry agrees lvltli recent experimental Fourier transform infrared spectroscopy asof atomic carbon C ( , P , ) with unsaturated C(,P,) + CIH:(XIC-,) + CjH2-+ s i g ~ l m e n t sbased o n isotope s u b s t ~ t u t i o n hyclrocarbons ( 3 ) , a n d co-~vorkers ~ m ~ l e m e n t ethis i ~ reaction class into pe~ - c , H ( x ' ~ , ) H('Sl:I) ( l a ) stuciies in argon niatrices (8).A seconci isomer, \~inylidenecarbene,HICCC, has lieric models of t h e dark molecular cloud c-C,H(X2B,) H('S1::) ( l b ) C , L symmetry a n d lies 134.9 kJ mol-I Department of Chemlstp~,Unlversty of Callfornla, Berket hi^ systelll represellts [he prototyL,e reac. a h o \ ~ epropargylene. Its enthalpy of formalei!, CA 94720, USA, and Chemlca Sc~encesD~\!is~on, National Berkeley,CA tion of ubiclultous illterstellar carbon atoms tion AH; ( 0 K ) = 678.6 kJ mol-' ia in 94720, USA. lvith the sinlplest ~ul~saturated hydrocarbon excellent agreement with a n experimen-
79 XL'O\'EhlBER 1996