On the Possible Influence of Extraterrestrial Volatiles on Earth's

ELSEVIER Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51

On the possible in¯uence of extraterrestrial volatiles on Earth's climate and the origin of the oceans

David Deming * School of Geology and Geophysics, University of Oklahoma, Norman, OK 73019, USA Received 16 October 1997; revised version received 17 August 1998; accepted 24 August 1998

Abstract

A consideration of observational and circumstantial evidence suggests that Earth may be subject to high in¯ux rates (1011 ±1012 kg=yr) of extraterrestrial-sourced volatile elements (carbon, hydrogen, oxygen, nitrogen) derived from comets or other primitive solar-system material. The total extraterrestrial in¯ux rate may be four to ®ve orders of magnitude greater than previously thought, large enough to account for today's total near-surface inventories of water and carbon. The possibility of high rates of extraterrestrial volatile-accretion suggests a new climatic paradigm wherein Earth's surface temperature is in¯uenced by con¯icting internal and external processes. A variable in¯ux of volatile elements tends to warm the Earth, while terrestrial processes cool the planet by absorbing these gasses at a more uniform rate. Variations in extraterrestrial in¯ux rates may explain the variation of sea level and mean global temperature over geologic time, as well as some types of climate change, the occurrence of the Pleistocene ice ages, and the asymmetry of the Phanerozoic climate record (sudden warmings, slow coolings). The extraterrestrial in¯ux rate may also act as the pacemaker of terrestrial evolution, at times leading to mass extinctions through climatic shifts induced by changes in accretion rates with concomitant disruptions of the carbon and nitrogen cycles. Life on Earth may be balanced precariously between cosmic processes which deliver an intermittent stream of life-sustaining volatiles from the outer solar system or beyond, and biological and tectonic processes which remove these same volatiles from the atmosphere by sequestering water and carbon in the crust and mantle.  1999 Elsevier Science B.V. All rights reserved.

Keywords: paleoclimatology; climate; catastrophism; ice ages; mass extinctions; biologic evolution; comets

1. Introduction Eocene boundary about 55.5 Myr ago (Kennett and Stott, 1991; Thomas and Shackleton, 1996; Dickens Over the past ten to ®fteen years, our understand- et al., 1997). Simultaneous with this warming, an im- ing of Earth's climate has been challenged by new mense quantity of carbon greatly enriched in 12Cwas data which cannot easily be explained by existing rapidly added to the combined ocean±atmosphere in- paradigms. Cores collected as part of the Deep Sea organic carbon reservoir. The amount of carbon that Drilling Program document that a remarkably large was added is much greater than can be obtained from and abrupt warming occurred near the Paleocene± any conventional terrestrial reservoir (Dickens et al., 1997). Ł Tel.: C1-405-325-6304; Fax: C1-405-325-3140; E-mail: Analyses of ice-cores and ocean-¯oor sedi- [email protected] ments show that repeatedly during the last ice age

Earth's climate underwent frequent, large and abrupt going extraterrestrial accretion. The extraterrestrial changes that were synchronous worldwide (Dans- volatile-accretion (ETV) hypothesis also postulates gaard et al., 1993; Grootes et al., 1993; Blanchon and that the accretion of volatile elements may have had Shaw, 1995; Nicholls et al., 1996, p. 177; Broecker, a signi®cant effect on Earth's climate over geologic 1997a,b). The ultimate cause of such abrupt shifts time, and thus in¯uenced organic evolution. With the in mean global temperature is unknown. It was pre- exception of observational evidence for the accretion viously thought that these rapid changes in climate of small, water-rich comets (Frank and Sigwarth, were con®ned to the North Atlantic and Arctic, and 1993, 1997a,b, 1997c,d), the evidence for an ETV thus could be explained by changes in Atlantic circu- hypothesis is largely circumstantial and subject to in- lation. However, the discovery that these climatic terpretation. Taken by itself, each line of evidence is events were worldwide makes them much more weak and inconclusive; it is the sum of the evidence dif®cult to explain (see discussions by Broecker, that is compelling. Phenomena which are apparently 1997a,b). unrelated and dif®cult to explain with existing the- New results have also cast doubt on whether ories fall neatly into place when the possibility of or not the Milankovitch theory adequately explains a previously unsuspected astronomical mechanism is the 100,000-yr cycle of Pleistocene glaciations. Pre- considered. cise radioactive-dating of a calcite vein from Devil's In the following sections, I ®rst review in moder- Hole, Nevada, shows that about 140,000 yr ago an ate detail two examples of climatic mysteries which interglacial period may have already been under- suggest that our current understanding of Earth's cli- way, even though solar insolation was at a minimum mate may be incomplete. The absence of suitable (Winograd et al., 1992). Analyses of proxy climate hypotheses thus implies the necessity of entertaining records from cores collected in the western Paci®c a new hypothesis. I then present the ETV hypothe- Ocean have found that the inclination of Earth's or- sis itself, broadly painted in general terms, with the bit may drive the Pleistocene glacial cycles, even understanding that what is being proposed is not to though inclination has no signi®cant effect on solar be interpreted as dogma, but as working hypothe- insolation (Muller and MacDonald, 1997a,b). sis subject to being tested and possibly disproven. Over the same time period that new data have I review evidence that supports the hypothesis, and challenged existing climatic paradigms, quantitative ®nally discuss and speculate on some of the possible estimates have shown that volcanic outgassing rates consequences. are too low to explain the origin of the oceans and the crustal carbon inventory (Bebout, 1995). Water and carbon are being carried down into the mantle by 2. Two climatic mysteries subduction much faster than they are being released to the surface by outgassing. Simultaneously, miner- In this section, I brie¯y review the existence of alogic studies have concluded that subduction may two climatic phenomena for which we either have carry vast amounts of water down into the mantle no viable hypothetical explanations, or for which ex- where it can be stored in the form of dense hydrous isting hypotheses have large dif®culties. The relative magnesium silicates (DHMS) (Smyth, 1994; Bose importance of some of these dif®culties is subject and Navrotsky, 1998). to interpretation and may be extremely controversial. It is entirely possible that these separate and ap- However, there can be no doubt that there exist some parently unrelated phenomena could be attributable terrestrial climatic phenomena for which data may to one or more terrestrial mechanisms which are con¯ict with current theories in some respects. unknown or poorly understood at the present time. However, it is also possible that these unsolved prob- 2.1. The Pleistocene ice ages lems could be explained through a new and unifying theory. The purpose of this paper is to suggest as a For the last 60 to 70 million years, the Earth has working hypothesis that Earth's supply of volatile been cooling, especially at high latitudes (Fig. 1). elements may have largely originated from on- The culmination of this cooling trend has been the D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 35

ing trend which began about 60 million years ago. The Milankovitch theory only seeks to explain the relative timing of glacial±interglacial episodes. Even in that respect, however, there are dif®culties. Al- though there can be no doubt that variations in solar insolation affect Earth's climate, the 100,000-yr insolation cycle is out of phase with the 100,000-yr climate cycle (Imbrie et al., 1993a). Because Earth's climate system is complex, phase problems are not necessarily a fatal weakness. For example, Imbrie et al. (1993a) have argued that phase lags are a natural consequence of the thermal inertia of the ice sheets. However, there are other problems. Milankovitch theory predicts insolation cycles of 23,000, 41,000, and 100,000 yr. The largest temperature changes have a 100,000-yr periodicity, but the 100,000-yr Milankovitch cycle exhibits the smallest insolation variation. In fact, changes in insolation associated with the 100,000-yr Milankovitch cycle (0.2%) are too small by a factor of ten to explain planetary temperature changes (Frakes et al., 1992, p. 141). The 100,000-yr insolation cycle may be ampli®ed by terrestrial feedbacks. However, we then encounter the Fig. 1. Estimated mean global temperature and sea level for logical problem of explaining why the larger 23,000- Phanerozoic times. Sea level curves are from Vail et al. (1977) and 41,000-yr insolation cycles are not similarly and Hallam (1992). Note in¯ections in sea-level curves at both ampli®ed by terrestrial processes. the Permian±Triassic and Cretaceous±Cenozoic boundaries (after The Milankovitch theory may also suffer from Frakes et al., 1992). causation problems. Isotopic data from Devil's Hole, Nevada, and the Vostok Ice Core from the Antarctic, occurrence of the Pleistocene ice ages, wherein mean show that prior to the penultimate glacial termination planetary temperatures drop periodically by 5ëC or about 140,000 yr ago, temperature began to increase more and large areas at high latitudes are covered by when insolation was decreasing. Temperature con- vast ice sheets. For the last 600,000 yr, ice ages have tinued to increase for 10,000 yr while insolation occurred in a 100,000-yr cycle (Frakes et al., 1992, decreased (Fig. 2). If variations in insolation control p. 138). The prevailing theory which explains this planetary temperature, as the Milankovitch theory glacial cycle is the Milankovitch theory of variations predicts, it is dif®cult to understand how tempera- in insolation related to changes in Earth's orbital ture could have begun rising before solar insolation parameters (Milankovitch, 1930; Hays et al., 1976; increased. It is also dif®cult to avoid this logical Imbrie et al., 1992, 1993a,b). As more detailed and contradiction by supposing that the Devil's Hole and precise climatic data have become available, how- Vostok records are not representative of global cli- ever, certain inadequacies in the Milankovitch theory matic conditions, but rather record local warming have become apparent (see Muller and MacDonald, events. The simultaneous initiation of local warming 1995, for a succinct review). trends, each lasting 10,000 yr, at sites in different First, the Milankovitch theory does not explain the hemispheres, would be a profound coincidence. It cooling over geologic time which is the actual cause is more parsimonious to interpret the Devil's Hole of the ice ages. The Pleistocene glaciations did not and Vostok records as the common re¯ection of a suddenly spring into existence without antecedent. global trend (Winograd et al., 1992, 1997; Shaffer They are the logical culmination of a long-term cool- et al., 1996; Edwards et al., 1997). Another causa- 36 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51

Fig. 2. June insolation at 60ëN and isotopic variations from the Vostok Ice Core, Devil's Hole vein calcite (DH-11), and marine cores (SPECMAP). Solid vertical lines represent time of glacial terminations as inferred from Vostok and Devil's Hole records. Dashed vertical lines represent time of terminations as inferred from the SPECMAP record (after Winograd et al., 1992, p. 257). tion problem is that isotopic data from both Devil's modulate a process such as vulcanism into releas- Hole (Winograd et al., 1992, 1997) and SPECMAP ing CO2 at regular intervals. To obtain a rhythmic (Imbrie et al., 1984) show that a glacial termination forcing periodicity, we need to look to an astronom- (Termination V) occurred about 420,000 yr ago even ical mechanism. The only known forcing function though there was no signi®cant increase in insola- with a 100,000-yr periodicity is a Milankovitch in- tion. solation cycle. Paleoclimatologists are thus led to While the 100,000-yr insolation change is out of the awkward conclusion that greenhouse-gas con- phase with climate change, concentrations of green- centrations follow temperatures, reversing the usual house gasses are both correlated with temperature thermophysical relationship. and in phase. Raynaud et al. (1993) described the The lack of an alternative hypothesis to explain overall correlation between temperature and CO2 the correlation between temperature and greenhouse- concentrations inferred from analysis of Antarc- gas concentrations forces con¯icting data to be ig- tic ice cores as `remarkable'. Most conventional nored. For example, although feedback mechanisms climatic theories explain the correlation of green- undoubtedly operate to some degree, there is some house-gas concentrations and temperature through question as to whether or not they can fully ex- feedback mechanisms that amplify Milankovitch in- plain the high in-phase correlation of CO2 with tem- solation variations (e.g., see discussion by Raynaud perature changes inferred from oxygen isotopes. If et al., 1993), reversing the normal cause and effect increases in CO2 concentrations are caused by tem- relationship between temperature and greenhouse- perature changes, increases in CO2 concentrations gas concentration. This reversal is an awkward con- should lag temperature increases, yet they do not. sequence of forcing data to ®t theory. Planetary During the penultimate deglaciation (about 140,000 temperatures for the last 600,000 yr have exhib- yr ago), increases in CO2 concentrations unambigu- ited a 100,000-yr periodicity. If the periodicity is ously preceded local temperature increases (Raynaud attributed to the effect of greenhouse-gas changes, et al., 1993). Clearly, these data do not support the then there must be a terrestrial mechanism that idea that greenhouse-gas concentrations follow tem- produces changes in greenhouse-gas concentrations perature. with a 100,000-yr periodicity. However, it is dif®cult Estimates of paleotemperatures from low-latitude to imagine any terrestrial mechanism which might sites in Jamaica and Australia (e.g., Goodfriend D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 37 and Mitterer, 1993; Miller et al., 1997) have found 10,000 yr. There was a simultaneous disruption of that Late Pleistocene land temperatures were 8±9ëC the planetary carbon-cycle, with the concentration of colder than during the Holocene. It is thus apparent 13C decreasing rapidly. Carbonate and organic matter that the Pleistocene ice ages represented a profound show a δ13C change of about 2.5½ over 104 yr, change in the thermal state of the entire Earth; they followed by an exponential return to initial values were not merely changes in the manner in which over about 2 ð 105 yr. The carbon excursion was heat was distributed over the surface of the planet. global in its nature, with isotope anomalies occur- The ice ages cannot be explained solely through ring in both deep and surface ocean waters as well changes in terrestrial mechanisms such as ocean cur- as on continents. The remarkable magnitude, tim- rents or winds which merely redistribute the sun's ing, and global nature of the δ13C excursion imply heat. Planetary-wide changes in temperature imply that an immense quantity of carbon greatly enriched fundamental changes in either the amount of solar in 12C was rapidly added to the combined ocean± energy reaching Earth or the radiative balance of the atmosphere inorganic carbon reservoir (Kennett and planet. Stott, 1991, p. 225; Dickens et al., 1997, p. 259). A vexing question to which we have no good an- The temperature and carbon anomalies were accom- swer is this: why do ice ages end? Why did the fun- panied by a mass extinction of benthic foraminifera; damental thermal balance of the planet in Pleistocene the extinction was unique in its global synchronicity, times change abruptly every 100,000 yr? There are extent, and rapidity (Thomas and Shackleton, 1996, positive feedbacks associated with planetary cooling p. 403). which should have led to further cooling. The growth The Paleocene±Eocene anomaly can not be ex- of ice sheets increases the mean albedo of the planet plained by the addition of CO2 from volcanic and reduces the amount of solar radiation absorbed. sources, because mantle CO2 is not suf®ciently en- Greenhouse warming is also reduced by colder tem- riched in 12C. The anomaly also can not be explained peratures. Cold air contains less water vapor than by a transfer of terrestrial biomass, because the ter- warm air, and cooler oceans would have retained restrial biomass is not large enough (Dickens et al., CO2 at higher concentrations. It is thus not surpris- 1997). Dickens et al. (1997) recently proposed that ing that the record of the last four glaciations (Fig. 2) the isotopic anomalies near the Paleocene±Eocene is generally one of decreasing temperature over time. boundary could be explained by the sudden release What is extraordinary is the sudden end of these ice of methane from hydrates in the upper few hundred ages. If, as Winograd et al. (1992) wrote, ªorbitally meters of the oceanic crust. Although methane re- controlled variations in solar insolation were not a lease may have been an important factor in the rapid major factor in triggering deglaciationsº, then why climate change that occurred near the Paleocene± did warm interglacial episodes appear suddenly from Eocene boundary, there are some problems with the coldest extent of the Pleistocene ice ages? The this hypothesis as a sole explanation. Heat-conduc- 100,000-yr periodicity of the glacial cycle suggests tion theory shows that the rate at which a sudden an astronomical origin, yet the 100,000-yr insolation warming at the sea-¯oor surface propagates into the cycle is both out of phase with the climate cycle and subsurface and causes the release of methane de- too small to explain planetary temperature changes. creases exponentially with increasing time (Carslaw and Jaeger, 1959). It may thus be dif®cult to achieve 2.2. The Paleocene±Eocene boundary a methane release suf®ciently rapid to satisfy the constraints, depending upon the rate of warming, the The Paleocene±Eocene boundary (about 57±58 distribution of clathrates, and the thermal proper- Ma) is marked by a unique climatic and biotic ties of the oceanic crust. The hypothesis also leaves disturbance (Kennett and Stott, 1991; Thomas and unanswered the cause of the warming which released Shackleton, 1996; Dickens et al., 1997). Oxygen iso- the methane. The ocean ¯oor can be warmed by a tope data indicate that in the Late Paleocene (about reversal in the pattern of deep-ocean circulation, but 55 Ma), temperatures at high-latitude locations and then we are left with the question of what caused the in the deep ocean increased by 4±6ëC in less than change in oceanic circulation. 38 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51

rates orders of magnitude higher than previously suspected. Volatile accretion through geologic time has resulted in the accumulation of signi®cant stores of water and carbon near the Earth's surface. Volatile accretion is also postulated to affect the radiative balance of the planet, and thus climate and organic evolution. The extraterrestrial-volatile-accretion (ETV) hypothesis has the following tenets: (1) Earth accretes substantial amounts of volatile elements (carbon, hydrogen, oxygen, nitrogen) from extraterrestrial sources. These volatiles may enter Earth's atmosphere in the form of compounds such as H2O, CO2,CH4,NH3,CO,andNO2, thought to exist in comets. Other elements may oxidize during atmospheric entry (e.g., carbon to carbon dioxide) or undergo complex chemical reactions. The exact composition and in¯ux rate is unknown; however, the in¯ux rate is high enough to exert a climatic Fig. 3. Carbon-isotope curve for a core from the Carnie Alps, in¯uence, and therefore orders of magnitude higher Austria, encompassing the Permian±Triassic boundary. Core cov- than previously suspected. The nature of the climatic ers about 50 m of section; alternating black-and-white bars rep- effect is to increase the mean planetary temperature, resent about 110,000 yr of time per bar. Arrows indicate sites of elevated iridium (Ir) concentrations (after Rampino and Hag- this is probably accomplished primarily through the gerty, 1996, p. 16). accumulation of CO2 in the atmosphere. (2) The in¯ux rate may be modulated by one or more periodicities related to astronomical factors. The Paleocene±Eocene boundary is unique in that The in¯ux rate may also be highly variable, as has the magnitude and rapid timing of the 12C excur- been observed for more conventional objects such sion exclude the possibility that any conventional as meteors, whose in¯ux rate at times may be as mechanism in the global carbon-cycle was a primary much as 100,000 times the background rate. Sudden cause of the carbon-cycle disruption (J. Dickens, and large in¯uxes of extraterrestrial volatiles may pers. commun., 1997). However, the events which disrupt planetary biological and geochemical cycles occurred at the Paleocene±Eocene boundary were and leave isotopic markers in the geologic record. not unique in their character. The geologic record (3) The prevailing assumption of balanced hydro- shows a repeated pattern over Phanerozoic times of logic and carbon cycles is false. The rate at which mass extinctions coinciding with iridium and sta- carbon and water are ®xed in the solid Earth ex- ble-isotope anomalies (see review by Rampino and ceeds the rate at which these elements return to Haggerty, 1996). Major biostratigraphic boundaries the atmosphere, oceans, and biosphere through pro- commonly exhibit an increase in the concentration cesses such as volcanic outgassing. Fixation of car- of iridium with simultaneous or nearly simultaneous bon through terrestrial processes such as weathering, carbon and oxygen isotopic anomalies (Fig. 3). The sedimentation, and subduction, tends to cool the usual interpretation of the oxygen anomalies is a Earth as the concentration of atmospheric CO2 drops global warming of the order of 1 to 10ëC. (e.g., Raymo et al., 1988; Berner, 1990; Raymo and Ruddiman, 1992). Terrestrial processes are thus in- herently inimical to life, and the Earth's natural state 3. The ETV hypothesis in the absence of ongoing volatile accretion may be an ice-covered planet. The con¯ict of accretionary As a working hypothesis, I suggest that Earth is and terrestrial processes leads to an asymmetry in subject to ongoing accretion of volatile elements at the climatic record, as warming accretionary events D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 39 tend to be rapid, abrupt, and discontinuous, while 4. Supporting evidence cooling terrestrial processes tend to be slower and more uniform in their temporal character. 4.1. The small-comet hypothesis (4) Unusually high or low volatile-accretion rates may be responsible for high extinction rates and The small-comet hypothesis (Frank et al., 1986) moderate the evolution of life on Earth. Low volatile- proposes that Earth is currently accreting water at accretion rates produce biotic stresses by leading a rate of (0.2±1.0) ð 1012 kg=yr from an in¯ux of to cold and dry climates. High volatile-accretion small [mass (0.2±1.0) ð 105 kg, diameter about 4±6 rates produce more equable warm and humid cli- m], comet-like objects. Small comets are believed mates, during which life may ¯ourish. Catastrophic to be composed primarily of water, but a carbon accretion events such as likely happened at the mantle is required to avoid vaporization in inter- Cretaceous±Tertiary border may not be isolated planetary space. Putative estimates of carbon-mantle events, but extreme end points on a continuum. The density and thickness made by Frank and Sigwarth extraterrestrial-accretion rate may act as the pace- (1993) imply that small comets are probably about maker of biologic evolution by both sustaining and 90% water, 10% carbon. The extent to which these destroying life. hypothetical bodies may contain minor amounts of (5) When the Earth and solar system were other volatile elements is unknown. The current rate younger, accumulation events were likely larger in of water accretion from small comets is suf®cient magnitude or more frequent, or both. As geologic to increase mean global sea level by 2±10 mm in time has progressed, the extraterrestrial in¯ux has 3600 yr, and account for the entire mass of Earth's likely lessened as the source reservoir has dimin- oceans in 1.4±7.0 Ga. The inferred small-comet ac- ished. However, the chemical (e.g., silicate weath- cretion rate is 4 to 5 orders of magnitude higher ering), mechanical (e.g., subduction), and biological than generally accepted for the total accretion rate of processes (e.g., formation of limestones) which ®x all extraterrestrial material (Tuncel and Zoller, 1987; carbon and water in Earth's crust and mantle have Peucker-Ehrenbrink, 1996) continued with little abatement. Thus the overall Evidence for the existence of small comets comes trend over geologic time has been for the surface of largely from satellite observations of Earth's atmo- the Earth to cool. sphere. Data include global images of Earth's ul- (6) The source of accreted volatiles is unspeci®ed traviolet dayglow, detection of atomic-oxygen trails and unknown at the present time. The most likely and OH emissions from small comets, microwave ra- candidates for the extraterrestrial reservoir in which diometer observations of transient water-bursts in the accreted material originates are the classical Oort atmosphere, and visual sightings from ground-based cloud and the massive inner Oort cloud (Hills, 1981), telescopes (Frank and Sigwarth, 1993, 1997a,b, which may contain up to 100 times as many comets 1997c,d). as the classical Oort cloud (Weissman, 1986, p. Following its introduction in 1986, the small- 219). Sources outside the solar system are also con- comet hypothesis proved extremely controversial. ceivable. The mechanism which delivers volatiles to The original publication in Geophysical Research Earth is similarly unknown and unspeci®ed. Several Letters (Frank et al., 1986) unleashed a ¯ood of crit- mechanisms are possible. These include an unseen icism (e.g., see Davis, 1986; Donahue, 1986; Chubb, solar companion (the Nemesis of Davis et al., 1984), 1986; Dessler, 1991). Both of the reviewers of the a dark planet (Frank, 1990), random passing stars, original paper recommended against publication, one encounters with giant molecular clouds (Weissman, of them noting that ªif this is correct, we would have 1986), and galactic tides (Matese and Whitmire, to burn half the contents of the libraries in the physi- 1996). cal sciencesº (Frank, 1990, p. 23). The small-comet hypothesis has recently received renewed attention due to apparent substantiation from a new generation of satellites (Frank and Sig- warth, 1997a,b, 1997c,d). However, these new data 40 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 have been followed by a new generation of criti- of Precambrian age is about the same as that in cisms. Parks et al. (1997, 1998) supported Dessler's younger sedimentary rocks. (1991) hypothesis that the data interpreted by Frank Rubey (1951) reasoned that carbon has been and Sigwarth (1993, 1997a,b, 1997c,d) to be small added to the atmosphere and oceans continuously comets are instrumental artifacts. However, Frank through geologic time. The amount of carbon tied and Sigwarth (1998) have disputed the interpreta- up in rocks is 600 times the amount of carbon tions of Parks et al. (1997). Grier and McEwen present in today's atmosphere, hydrosphere, and bio- (1997) found no visual evidence for small-comet sphere combined. As carbon is continuously ex- impacts on the moon, adding to criticisms by ear- tracted from the atmosphere±ocean system by sedi- lier workers (Davis, 1986; Nakamura et al., 1986) mentation, Rubey (1951) pointed out that if a contin- that small-comet impacts on the moon were unde- uous rate of supply were not present, brucite would tected by seismometers left by Apollo astronauts. replace calcite as a common marine sediment. Yet Rizk and Dessler (1997) argued that the collision of the geologic record contains no brucite deposits. small comets with Earth's atmosphere should cre- Rubey (1951) concluded that the geologic record ate intermittent appearances of bright, rapidly mov- strongly indicated that Earth's oceans and crustal ing points of light. Because such events have not carbon inventory formed as the result of slow ac- been observed, Rizk and Dessler (1997) concluded cumulation over geologic time from a more-or-less small comets do not exist. Summers et al. (1997) continuous and gradual supply mechanism. The only found what appeared to anomalously high levels such mechanism known to Rubey (1951) was vol- of water near altitudes of 65 to 70 km in the at- canic outgassing. Earth scientists have subsequently mosphere, suggesting the possibility of an external accepted with little question that the Earth's oceans source. However, Hannegan et al. (1998) concluded and its near-surface inventory of volatiles such as that the amount of water in the stratosphere was 100 carbon have originated from volcanic outgassing. times less than that consistent with the small-comet However, Rubey (1951) never measured the rate hypothesis. at which water and carbon are added to the ter- Criticisms of the small-comet hypothesis are restrial surface environment through volcanic out- abundant, but it is not clear at the present time if gassing. Any attempt by Rubey (1951) to make such the small-comet hypothesis is faulty or if the criti- estimates would have been badly ¯awed, because he cisms instead are based upon faulty assumptions. It was unaware of the process of subduction. Subduction seems likely that the small-comet hypothesis shall carries both water and carbon down into the mantle remain controversial for some time to come. at signi®cant rates, and it is questionable if the water and carbon thus removed from the near-surface 4.2. Origin of water and carbon inventories environment is fully returned through outgassing. For the past 15 years or so, quantitative estimates In a classic paper on the origin of the oceans, have been made which compare the rate at which Rubey (1951) pointed out that volatile compounds water and carbon are released by volcanic outgassing found near the Earth's surface such as water and to their uptake by subduction (see review by Bebout, carbon dioxide are much too abundant to have 1995). It turns out that the net rate at which water formed through the process of rock weathering. is added to the hydrosphere by volcanic outgassing Rubey (1951) also concluded that all of the volatile is only 5±15% of the rate at which it is carried compounds found near the surface today could not down into the mantle by subduction. Similarly, the have been present early in the Earth's history in the net rate at which carbon is added to the atmosphere form of a `dense primitive atmosphere'. High atmo- by volcanic outgassing is only 10±44% of the rate spheric concentrations of carbon dioxide, nitrogen, at which it is lost to the mantle (Table 1). Based and hydrogen sulfate in contact with sea water would upon all existing estimates and studies, volcanic have led to a highly acidic ocean with the concomi- outgassing is a grossly inadequate mechanism to tant deposition of large quantities of carbonate rocks. explain the abundant existence of water and carbon However, the percentage of limestone in sediments at and near the Earth's surface. D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 41

Table 1 inventories can be estimated as: Rates at which water and carbon are consumed by subduction Supply Rate D compared to production from volcanic outgassing (data from Bebout, 1995) Present Inventory C Depletion Rate Compound Consumption rate Production rate Age Earth (1010 kg=yr) as % consumption where `depletion rate' refers to the net depletion rate Water 91±194 5±15 (subduction minus outgassing). Based upon current Carbon 6±28 10±44 estimates from several studies (Table 1), net depletion rates for water and carbon are in the range of (0.8±1.8) ð 1012 kg=yr and (0.3±2.5) ð 1011 kg=yr, If near-surface supplies of water and carbon did respectively. Estimating present water and carbon in- not come from volcanic outgassing, where did they ventories to be 1.37 ð 1021 kg and 1.0 ð 1020 kg, originate? There are at least three possible explana- respectively, and the age of the Earth to be the age tions. First, it is eminently possible that substantial of the oldest known rock (4.0 ð 109 yr), the water amounts of water and carbon return to the surface and carbon supply rates necessary to explain current through processes other than volcanic outgassing. inventories are in the range of (1±2) ð 1012 kg=yr One possibility is updip transport back up a sub- and (0.6±3) ð 1011 kg=yr, respectively. These esti- ducting wedge. Sea¯oor ¯uid venting occurs along mates are not signi®cantly affected by the assumed trenches, and studies of accretionary complexes doc- amount of volatiles present early in the Earth's his- ument updip ¯uid ¯ow (Fisher, 1996). On the other tory, because the calculated supply rate is largely hand, the estimates of water loss rates cited in Ta- determined by the depletion rate. ble 1 are very conservative because they only include Frank and Sigwarth (1993) estimated water was water which is mineralogically bound. Pore ¯uids, being added to Earth from small comets at a rate of which account for 90% of subducted water, are not (0.2±1.0) ð 1012 kg=yr. Based upon putative esti- included. It is therefore possible that all accounts mates of carbon-mantle density and thickness made of sea¯oor venting and other evidence for shallow by Frank and Sigwarth (1993), small-comet carbon return near subduction zones could be accounted for content is probably equivalent to about 10% of the by the return of pore ¯uids. total water mass. Thus small-comet carbon-accretion A second possibility is that the rate of volcanic rates are in the range of (0.2±1.0) ð 1011 kg=yr. outgassing was higher early in the Earth's history. In consideration of the uncertainties involved in as- This is not really an attractive hypothesis, however. suming accretion, subduction, and outgassing rates High outgassing rates would have resulted in the constant over geologic time, these numbers are very `dense primitive atmosphere' which Rubey (1951) close to those which satisfy the geologic constraints. showed was geochemically impossible. It is also If water and carbon are being carried down into probable that high levels of volcanic activity on a the mantle and not returned, there should be substan- young and hotter Earth would have been matched by tial amounts of these elements stored in the Earth's higher subduction rates. interior. Presumably, most storage would take place A third possibility is the existence of an external in the upper mantle, although the degree to which source of supply, perhaps similar to the small comets the lower and upper mantle mix is poorly known. inferred by Frank et al. (1986). This hypothesis Assuming that the rates cited in Table 1 have been can be put to a test of logical consistency. That is, constant over geologic time, the net amount of water how does the estimated rate of water and carbon carried down into the mantle over the last 4:0ð109 yr accretion inferred from satellite data compare to should be in the range of (3±7) ð 1021 kg, two to ®ve the rate implied by geological data as necessary to times the mass of today's oceans (1:37 ð 1021 kg). create and maintain Earth's near-surface inventories Ten years ago, there was little to no evidence for sub- of water and carbon? If it is assumed that the Earth stantial amounts of water storage in the upper mantle formed without water and carbon at the surface, (Ahrens, 1989). However, in the past few years it has then the supply rate necessary to explain existing become apparent that large amounts of water may be 42 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 stored in the upper mantle in the form of dense hy- 4.4. Correlations between sea-level trends and mass drous magnesium silicates (DHMS). One of the most extinctions important minerals in this group is wadsleyite, which maycontainupto3.3%H2O by weight (Smyth, The two largest mass extinctions of Phanerozoic 1994). Smyth (1994) estimated that if the mantle be- times (Permian±Triassic and Cretaceous±Tertiary) tween the depths of 400 and 525 km were composed were essentially synchronous with in¯ections in eu- of 60% fully hydrated wadsleyite, the amount of wa- static sea-level curves (Fig. 1). Dating of melt rock ter contained therein would be ªmore than four times from the Chicxulbu Crater (Swisher et al., 1992) has the amount of H2O currently in the Earth's hydro- now established at a high level of probability that the sphereº. More recently, Bose and Navrotsky (1998) mass extinction which occurred at the Cretaceous± showed ªthere is no barrier to subducting substantial Tertiary boundary was caused by a catastrophic ac- amounts of water to depths of 400±600 km in colder cretion event as suggested by Alvarez et al. (1980). slabs, since the slab can remain in the stability ®eld There is also substantial evidence that the Permian± of hydrous phases throughout its descentº. Other ev- Triassic extinction event was related to extraterres- idence for the presence of water in the upper mantle trial accretion. Although complete sections which is provided by seismic tomography (Nolet and Ziel- encompass the Permian±Triassic boundary are rela- huis, 1994, p. 15,816), excess helium in groundwater tively scarce (Bice et al., 1992), iridium anomalies (Torgersen et al., 1995), and deep-focus earthquakes at the Permian±Triassic boundary have been found (Meade and Jeanloz, 1991). at 10 different locations (Rampino and Haggerty, 1996, p. 15). Shocked quartz, diagnostic of im- 4.3. Correlations between sea level and temperature pact ejecta, has also been found near the Permian± Triassic boundary at a site in northern Tuscany, Italy If the ETV hypothesis is correct, geologic peri- (Bice et al., 1992). ods characterized by high rates of volatile accretion Temporal correlations between changes in sea should be times of relative warmth during which sea level, extraterrestrial accretion, and biologic evolu- level rises; while low accretion rates should result tion, suggest the probability that these phenomena in falling sea level and cooling. This correlation is are causally related. Conventional theories attribute present in the geologic record (Fig. 1). Each of the large eustatic changes in sea level to changes in the three eras of Phanerozoic time (¾570 Ma±present) volume of the ocean basins that result from changes show general correlations between temperature and in the rate of sea-¯oor spreading. If we are to accept sea level (Frakes et al., 1992, p. 194; Hallam, 1992, this theory, however, then we have to believe that p. 158). The correlation is not perfect for at least catastrophic accretion events coincidentally occurred two reasons. Both climate and sea level are affected at the very moments in geologic time when sea-¯oor by numerous processes and factors in addition to spreading rates changed. This seems implausible. It hypothetical changes in the rate of extraterrestrial ac- may be more parsimonious to postulate that the same cretion, and the Phanerozoic record of sea level and gravitational mechanism (e.g, galactic tides) which is temperature is partly qualitative and subject to future responsible for catastrophic meteoritic impacts may revision. For example, correlations between sea level simultaneously de¯ect large amounts of volatiles and global temperature exist partly because of glacial into Earth-crossing orbits. Rapidly changing terres- storage. However, there were signi®cant changes in trial climates might result from the accretion of these sea level over periods of Phanerozoic times when very different materials which have climatic effects little or no polar ice existed. Sea level dropped sig- over different time scales. The catastrophic impact ni®cantly from Late Cretaceous times (about 70 Ma) of one or more solid bodies could result in the at- through the Paleocene and Eocene (about 36 Ma), mosphere being loaded with smoke, dust, and sulfate yet there is little evidence for the existence of major aerosols. The amount of solar insolation reaching the ice caps or glacial deposition in ocean sediments Earth's surface would drop substantially for a period until the end of the Eocene (about 36 Ma) (Frakes et of time of 1±10 yr, and signi®cant cooling would al., 1992, p. 102). occur (Toon et al., 1997). However, as light-blocking D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 43 aerosols settled out of the atmosphere, the climatic (1995) in the form of a 100,000-yr periodicity in the effects of greenhouse gasses with longer residence ¯ux of extraterrestrial 3He to the Earth's sea ¯oor. times might dominate. The residence time of CO2 is However, Muller and MacDonald (1995, 1997a,b) about 50±200 yr, and high CO2 levels could lead to have never speci®ed the mechanism involved. Nei- a substantial greenhouse warming. This type of sce- ther has anyone else been able to propose a mech- nario could explain why oxygen isotope data show anism which could establish a causal link between that global warmings are commonly synchronous inclination and climate. One possibility is that me- with, or immediately postdate, several mass extinc- teoric materials may induce glacial episodes through tions, including the important Cretaceous±Tertiary aerosol cooling. However, there are two dif®culties and Permian±Triassic events (Rampino and Hag- with this hypothesis. First, glacial episodes do not gerty, 1996). require an explanation of this type. The Earth has been cooling at high latitudes for 60 million years, 4.5. The inclination cycle and glaciations are most parsimoniously interpreted as the logical consequence of whatever factors are re- Like orbital eccentricity, the inclination of Earth's sponsible for this long-term cooling trend. Once they orbit about the plane of the solar system has a begin, glacial episodes induce positive climatic feed- 100,000-yr periodicity. There is also a strong corre- backs that should result in further cooling. The high lation between the inclination of Earth's orbit and albedo of increased ice cover, the higher solubility of climate over the last 600,000 yr (Fig. 4; Muller CO2 in colder oceans, and decreased amounts of wa- and MacDonald, 1995, 1997a,b). Recent analyses of ter vapor in a colder atmosphere should all tend to re- ocean cores have found climate signals related to duce planetary temperature. As glaciations proceed, both eccentricity and inclination (Muller and Mac- ice volume increases. Thus, it is the sudden appear- Donald, 1997a,b). However, inclination appears to ance of interglacial episodes from the coldest extent be much more strongly related to climate. This is of the ice ages which is anomalous and requires an surprising, because the inclination of Earth's orbit explanation in the form of a strong external forcing. does not have a signi®cant effect on solar insolation. A second dif®culty with an aerosol-cooling hy- Muller and MacDonald (1995) suggested that the pothesis is that the accretion rate necessary to inclination of Earth's orbit may drive Pleistocene achieve global coolings of the magnitude associ- glacial cycles by modulating the accretion of ex- ated with Pleistocene glaciations (5ëC) is likely to be traterrestrial materials. Some con®rming evidence many orders of magnitude greater than the observed for a 100,000-yr periodicity in extraterrestrial accre- accretion rate of meteoric materials. If we assume tion rates has been found by Farley and Patterson that accreted material is as ef®cient in generating a

Fig. 4. Orbital inclination (solid line, lagged by 33 ka) and δ18O climate proxy (after Muller and MacDonald, 1995). 44 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 cooling effect as sulphate aerosols, an order-of-mag- orientation of the dust or other material whose accre- nitude calculation can be made of the extraterrestrial tion may be modulated by orbital inclination is un- ¯ux necessary to induce a 5ëC global cooling. known. There is no reason to postulate that the lowest The total anthropogenic sulfur-aerosol ¯ux today or highest accretion rates should occur during arbi- is 1:4 ð 1011 kg=yr, which is estimated to exert a trary values of inclination. Second, global ice volume total climatic forcing of 0.4 W=m2 with a factor of is itself not in phase with climate change. Winograd two uncertainty (Schimel et al., 1995, pp. 104, 113). et al. (1997) concluded that there are phase offsets The relationship between mean global temperature of thousands of years between proxies for global ice change .∆T / and radiative forcing .∆Q/ is volume (e.g., SPECMAP) and other climate proxies. It is possible to reconcile the spectral evidence ∆T D ½∆Q (1) for the Muller±MacDonald hypothesis with the crit- where ½ is the climate sensitivity parameter icisms listed above if the mechanism which affects (Cubasch and Cess, 1990, p. 77). Taking ½ D 0:3K climate is not the accretion of interplanetary dust, but m2 W1 (no climate feedbacks), the radiative forc- rather the accretion of extraterrestrial carbon. In this ing necessary to induce an ice-age cooling of 5ëC modi®ed form of the Muller±MacDonald hypothesis, is 17 W=m2. If a present-day anthropogenic sul- extraterrestrial accretion is modulated by orbital in- fur-aerosol ¯ux of 1:4 ð 1011 kg=yr exerts a climatic clination, but the climatic effect is not dust-induced forcing of 0.4 W=m2, then a simple linear extrap- cooling. Rather, the sudden appearance of inter- olation to the ¯ux required to produce a forcing of glacial episodes is attributed to temporary increases 17 W=m2 implies a necessary ¯ux of 5:8 ð 1012 in the carbon in¯ux rate with concomitant increases kg=yr. However, the accretion rate of conventional in atmospheric CO2 concentration and changes in meteoritic material averaged over tens of millions of the planetary radiative balance. As carbon-accretion years is well constrained to be in the neighborhood rates diminish, the planet once again cools as on- of 107 ±108 kg=yr (Tuncel and Zoller, 1987; Peucker- going terrestrial processes such as subduction and Ehrenbrink, 1996). Thus the accretion of extrater- sedimentation remove carbon from the surface faster restrial material appears to be four to ®ve orders of than it is introduced by accretion and outgassing. magnitude too small to exert a signi®cant effect on Earth's climate by direct aerosol cooling. Even if the 4.6. The temperature±volatiles conundrum linear extrapolation is not strictly justi®ed, the mass insuf®ciency is nevertheless enormous. If glaciations Standard planetary formation models show that are induced by the accretion of extraterrestrial dust, the Earth likely formed without its present day it would seem to be necessary to invoke a profound inventories of water, nitrogen, and carbon. At ex- ampli®cation by some unknown or exotic process. pected temperatures of accretion, volatile tempera- Recent calculations by Kortenkamp and Dermott tures are suf®ciently high that water, nitrogen, and (1998) also indicate that the terrestrial accretion carbon remain in the gaseous phase and are not of interplanetary dust particles from the zodiacal incorporated in planet-forming planetesimals. This cloud, the tenuous cloud of dust which envelops problem is termed the ªtemperature±volatiles conun- the four terrestrial planets of the solar system, is drumº (Chyba et al., 1994, p. 13). The conundrum not controlled by the inclination of Earth's orbit, is that although our current understanding precludes but rather by its eccentricity. If dust accretion is Earth forming with volatile elements, the elements controlled by eccentricity, and not inclination, it may are here. The standard hypothesis which accounts be dif®cult to link correlations between inclination for the missing volatiles is an Archean late-accre- and climate with dust accretion. tionary bombardment of volatile-rich material (An- Orbital inclination does differ in phase by 33 ka ders, 1989). However, there appears to be no inde- (see Fig. 4) from global ice volume as inferred from pendent evidence that such an event occurred. the SPECMAP data (Imbrie et al., 1984). However, The ETV hypothesis proposes that the volatile in- this is not an impediment to the viability of the ¯ux rate has declined over geologic time, but perhaps Muller±MacDonald theory, for two reasons. First, the by one order of magnitude, instead of three (Chyba D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 45 and Sagan, 1992, p. 127). In either case, high rates 4.8. Cooling of the Earth of volatile accumulation subsequent to Earth's formation are necessary to explain present day volatile Oxygen isotope data from cherts (Fig. 5) show inventories. The ETV hypothesis merely suggests that Earth has cooled signi®cantly over geologic that the in¯ux rate has declined gradually instead of time, even though solar luminosity has increased precipitously. by approximately 30% (Knauth and Lowe, 1978). The temperature decrease is thought to be due to 4.7. Water in the solar system decreases in the concentration of CO2 and other greenhouse gasses (Rye et al., 1995). The magnitude The presence of water on Earth's moon (Nozette of the inferred cooling is large. Knauth and Clemens et al., 1996) and the planet Mercury (Slade et al., (1995) concluded the data are consistent with a warm 1992) implies the existence of a transport process (>45ëC) early Earth which has progressively cooled which delivers water to the inner solar system. Sur- to the present day mean planetary air-temperature face features on the planet Mars show evidence of of 15±16ëC. Knauth and Clemens (1995) also con- running water (Baker, 1982). Frank (1990) has pro- cluded that the temperature history of Earth has been posed that the presence of surface water on Mars highly variable, with both cool periods and major may be cyclic. Accreted water would be stored in global warming events. The warming episodes lasted growing polar ice caps until a short-lived period of about 100 Ma, and were characterized by mean rapid volatile accretion could warm the planet suf®- global temperature increases of 10 to 15ëC. ciently so as to melt the polar ice caps and allow for The most common objection to the warm early the presence of increased water vapor in the atmo- Earth inferred from isotopic studies of cherts is ev- sphere. Further warming would occur from positive idence for ice ages as long ago as Early Proterozoic feedbacks as water vapor is itself a greenhouse gas. times (2.3±2.5 Ga). However, it is possible that Eventually, much of the water in the atmosphere these ice ages did not occur, or were short-lived would be lost to space, and Mars would once again anomalies that occurred during periods of otherwise cool and begin to store accreted water at its poles high temperatures. The primary evidence for Pro- until the cycle was repeated. terozoic glaciations is the presence of diamictites,

Fig. 5. Range of oxygen isotope data from cherts (after Knauth and Lowe, 1978, p. 218). 46 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 poorly sorted sediments which have been interpreted modes is gradual (Frakes et al., 1992, p. 195). Exam- as glacial till. Rampino (1994) argued that most, if ples include the sudden warming at the Paleocene± not all, diamictites may have originated as the fallout Eocene boundary (Kennett and Stott, 1991), the of ballistic ejecta from the impact of large asteroids Pleistocene ice ages (Fig. 2), and interstadial warm- or comets. If this were true, it would explain the puz- ings (Nicholls et al., 1996, p. 177), all discussed zling occurrence of `glacial deposits' at low latitudes earlier. Other than the ETV hypothesis, there is no in close proximity to indicators of tropical climates theory known to the author which even attempts to such as carbonates, red beds, and evaporites (Evans explain this asymmetry. et al., 1997). 4.11. Noctilucent clouds 4.9. Abrupt climate change Some models of mesospheric noctilucent-cloud The occurrence of rapid climate changes unex- formation require the accretion of extraterrestrial plainable by any known terrestrial process suggests water at a rate of 1011±1012 kg=yr (Lebedinets and the possible presence of an unsuspected catastrophic Kurbanmuradov, 1992). mechanism. Recent analyses of ice core data show that interstadial transitions between warm and cold conditions during the last ice age were rapid, of the 5. Discussion order of decades (Grootes et al., 1993). For example, a reconstruction of climate records based on cores 5.1. Accretion rates from Greenland, Antarctica, and the North Atlantic, reveals what appear to be about twelve rapid climate Estimates of volatile accretion rates from the changes over approximately the last 100,000 yr. The small-comet hypothesis are four to ®ve orders of changes are large in magnitude, and their correlation magnitude greater than generally accepted estimates between sites implies that they were global. Most (107 ±108 kg=yr) of the total extraterrestrial in¯ux typically, warmings of 5 to 7ëC that take place in a rate (Tuncel and Zoller, 1987; Peucker-Ehrenbrink, few decades are followed by slow coolings (Dans- 1996). Conventional estimates of in¯ux rates are gaard et al., 1993; Grootes et al., 1993; Nicholls et largely derived by extrapolating minute tracer (e.g., al., 1996, p. 177). iridium, osmium) concentrations up through sev- It is not possible to explain such rapid variations eral orders of magnitude based on assumed chon- by changes in solar insolation related to changes in dritic abundances. However, the volatiles which enter Earth's orbit; these occur on a much longer time Earth's atmosphere may be derived from an entirely scale. It is also dif®cult to attribute rapid temperature different class of objects whose composition does changes to changes in terrestrial processes such as not resemble chondrites (Frank et al., 1987). The ice volume changes or deep-ocean circulation which designation `small comet' does not refer to a comet operate on characteristic time scales of several hun- which is small. Rather, the bodies inferred by Frank dred to several thousand years. Sur®cial changes in et al. (1986) are thought to represent an entirely new ocean circulation may have important climatic con- class of celestial objects which lack the dust and sequences (Broecker et al., 1990); however, we are other constituents of conventional large comets. For still left with the question of what triggers these example, recent observations of small-comet impacts changes. are inconsistent with the presence of dust (Frank and Sigwarth, 1997d). Therefore, the small-comet 4.10. Asymmetry of the climate record in¯ux rate cannot be estimated by assuming chon- dritic composition. Similarly, there is no reason to There is a striking and unexplained asymmetry to believe that methods such as satellite observations, the reconstructed terrestrial temperature record that visual sightings, and radio-meteor studies, which are appears across all time scales. Warming episodes suited to the detection of large bodies composed generally occur suddenly, while the onset of cool of rock and=or metal, would ordinarily be capable D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 47

3 of detecting low-density (0.1 g=cm ) small comets of SO2 exceeded a critical threshold of about 1000 which disintegrate at altitudes greater than 300 km ppm. and are thought to be composed almost entirely of water, with small amounts of carbon and perhaps 5.3. Climate change other cometary volatiles. Although the radiative effects of high-altitude wa- 5.2. Disruption of the nitrogen cycle ter vapor are unknown, it seems unlikely that direct addition of water vapor to Earth's atmosphere is ca- Terrestrial geochemical cycles are extremely com- pable of an inducing a perceptible greenhouse effect. plex, and a complete discussion of the consequences Even if accretion rates were very high for short pe- of disrupting them is beyond the scope of this paper. riods of time, the residence time of water vapor in However, it is worth noting that there may be sev- the troposphere is so short (11 days) that it would eral possible ways in which extraterrestrial accretion dif®cult to maintain an appreciable increase in hu- could perturb the nitrogen chemistry of Earth's atmo- midity. It is more probable that the importance of sphere, oceans, and biosphere, with possible climatic water accretion is in building and replenishing the rami®cations. It has been recognized for a long time oceans over long periods of time. that catastrophic impacts of the type that apparently The accretion of carbon in the form of carbon happened at the Cretaceous±Tertiary border would dioxide may be a more ef®cacious mechanism for create large amounts of nitric oxide by oxidation of effecting climate change. During the last glacial nitrogen in the atmosphere (Toon et al., 1997). The termination (18 to 11 ka), atmospheric CO2 concen- amounts of nitric oxide which can be created by this tration rose from 200 to 280 ppm. Considering the mechanism are very large, and may exceed the mass increased uptake of carbon by ocean surface wa- of the impacting object. For example, Turco et al. ters and the biosphere that probably occurred as at- (1982) calculated that the mass of nitric oxide added mospheric CO2 concentrations increased, Sundquist to Earth's atmosphere by the 1908 Tunguska event (1993) estimated that the total amount of carbon may have been up to six times the total mass of the necessary to explain the entire deglacial budget was impacting object. Environmental effects associated in the range of (6.5±9.5) ð 1014 kg. Averaged over with nitric-oxide formation include acidi®cation of 7000 yr, this is equivalent to an average ¯ux of (0.9± the near-surface terrestrial and ocean environments 1.4) ð 1011 kg=yr. The current carbon-accretion rate and depletion of the Earth's protective ozone layer from small comets is estimated to be (0.2±1.0) ð (Hsu and McKenzie, 1985; Toon et al., 1997). Less is 1011 kg=yr. known about the complexities of nitrogen-cycle dis- Even if the ends of the Pleistocene ice ages were ruptions. Hsu and McKenzie (1985) suggested that caused by increases in the in¯ux of extraterrestrial acidi®cation due to nitric-oxide formation caused re- carbon, it remains to be shown that the radiative forc- lease of CO2 from the ocean to the atmosphere and ing associated with the concomitant addition of CO2 led to global warming at the Cretaceous±Tertiary to the atmosphere is strong enough to initiate inter- boundary. It is also not clear how a sudden and glacial episodes. The total increase of CO2 in Earth's massive input of nitric oxide would affect the pro- atmosphere during the last termination was about 80 duction of nitrous oxide, a powerful and relatively ppm (Sundquist, 1993). Assuming a radiative forcing 2 long-lived greenhouse gas, at the other end of the ni- of 1:8 ð 10 W=m for each 1 ppm CO2 increase trogen cycle. If impacting bodies contain even trace (Schimel et al., 1995, p. 92), the total radiative forc- 2 amounts of sulfur, as most comets and asteroids do, ing due to CO2 at the last termination was 1.4 W=m . the chemistry of nitric oxide formation in the atmo- Applying Eq. 1 above, this is approximately strong sphere by impact-created shock waves and heating enough to raise mean planetary temperature by only may not be straightforward. For example, Muzio and 0.4ëC, too small by a factor of ten to explain an inter- Kramlich (1988) found that the amount of nitrous glacial warming of 5ëC. However, the Milankovitch oxide produced in fossil-fuel combustion increased trigger is even smaller. The estimated magnitude of by two orders of magnitude when the concentration insolation changes associated with changes in orbital 48 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 eccentricity is only š0.5 W=m2, about one third of Alvarez, L.W., Alvarez, W., Asaro, F., Michel, H.V., 1980. Ex- the radiative forcing due to the CO2 increase. In traterrestrial cause for the Cretaceous±Tertiary extinction: ex- the absence of climatic feedbacks, no mechanism by perimental results and theoretical interpretation. Science 208, 1095±1108. itself seems energetically suf®cient. Anders, E., 1989. Pre-biotic organic matter from comets and asteroids. Nature 342, 255±257. 5.4. Isotopic anomalies Baker, V.R., 1982. The channels of Mars. Univ. Texas Press, Austin, TX, 198 pp. In theory, it should be possible to test for the pres- Bebout, G.E., 1995. The impact of subduction-zone metamor- phism on mantle±ocean chemical cycling. Chem. Geol. 126, ence of volatile accretion and its possible in¯uence 191±218. on Earth's hydrologic and carbon cycles by look- Berner, R.A., 1990. Atmospheric carbon dioxide levels over ing for correlations between isotopic markers and Phanerozoic time. Science 249, 1382±1386. climatic events. For example, major carbon-isotope Bice, D.M., Newton, C.R., McCauley, S., Reiners, P.W., anomalies occur precisely at the Cretaceous±Tertiary McRoberts, C.A., 1992. Shocked quartz at the Triassic± Jurassic boundary in Italy. Science 255, 443±446. and Paleocene±Eocene boundaries (Boersma, 1984, Blanchon, P., Shaw, J., 1995. Reef drowning during the last p. 269). In practice, however, it is dif®cult to discrim- deglaciation: evidence for catastrophic sea-level rise and ice- inate between cause and effect because terrestrial sheet collapse. Geology 23, 4±8. climatic processes also lead to changes in oxygen Boersma, A., 1984. Campanian through paleocene paleotempera- and carbon isotopic ratios. A corollary is that the en- ture and carbon isotope sequence and the Cretaceous±Tertiary boundary in the Atlantic Ocean. In: Berggren, W.A., Van Cou- tire practice of evaluating terrestrial processes from vering, J.A. (Eds.), Catastrophes and Earth History. Princeton isotopic shifts may need to be reevaluated. Univ. Press, Princeton, NJ, pp. 247±277. Bose, K., Navrotsky, A., 1998. Thermochemistry and phase 5.5. Life on Earth equilibria of hydrous phases in the system MgO±SiO2±H2O: implications for volatile transport to the mantle. J. Geophys. Res. 103, 9713±9719. If the young Earth was hot and relatively inhos- Broecker, W.S., 1997a. Will our ride into the greenhouse future pitable to life, the Cambrian explosion of life forms be a smooth one? GSA Today 7 (5), 1±7. may have been a natural consequence of the gradual Broecker, W.S., 1997b. Thermohaline circulation, the Achilles subsidence of the volatile in¯ux rate. As the concen- Heel of our climate system: will man-made CO2 upset the tration of greenhouse gasses in Earth's atmosphere current balance? Science 278, 1582±1588. Broecker, W.S., Bond, G., Klas, M., 1990. A salt oscillator in the declined, the climate cooled, and an environment glacial Atlantic? 1. The concept. Paleoceanography 5, 469± more hospitable to life emerged. 477. As Frank (1990) has pointed out, it is dif®cult to Carslaw, H.S., Jaeger, J.C., 1959. Conduction of Heat in Solids, envision a mechanism which could deliver volatiles 2nd ed. Oxford Univ. Press, Oxford, 510 pp. to Earth in a manner consistent with the inferred Chubb, T.A., 1986. Comment on the paper ªOn the in¯ux of small comets into the Earth's upper atmosphere I. Observa- presence of small comets. A nearly unique set of cir- tionsº. Geophys. Res. Lett. 13, 1075±1077. cumstance is needed. If a regular delivery of volatiles Chyba, C., Sagan, C., 1992. Endogenous production, exogenous was necessary for the emergence and maintenance of delivery and impact-shock synthesis of organic molecules: an life on Earth, then life in the universe may be more inventory for the origins of life. Nature 355, 125±132. uncommon than previously thought (Frank, 1990). Chyba, C.F., Owen, T.C., Ip, W.-H., 1994. Impact delivery of volatiles and organic molecules to Earth. In: Gehrels, T. (Ed.), Instead of a robust process which mediates its envi- Hazards Due to Comets and Asteroids. Univ. Arizona Press, ronment (Lovelock, 1979), life may be a delicate and Tuscon, AZ, pp. 9±58. chance phenomenon that springs up for short periods Cubasch, U., Cess, R.D., 1990. Processes and modelling. In: of time under fortuitous circumstances. Houghton, J.T. et al. (Eds.), Climate Change, The IPCC Sci- enti®c Assessment. Cambridge Univ. Press, Cambridge, pp. 69±91. Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D., References Gundestrup, N.S., Hammer, C.U., Hvidbert, C.S., Steffensen, J.P., SveinbjoÈrnsdottir, A.E., Jouzel, J., Bond, G., 1993. Evi- Ahrens, T.J., 1989. Water storage in the mantle. Nature 342, dence for general instability of past climate from a 250-kyr 122±123. ice-core record. Nature 364, 218±220. D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 49

Davis, P.M., 1986. Comment on the letter ªOn the in¯ux of Grier, J.A., McEwen, A.S., 1997. The small-comet hypothesis: small comets into the Earth's upper atmosphereº. Geophys. an upper limit to the current impact rate on the Moon. Geo- Res. Lett. 13, 1181±1182. phys. Res. Lett. 24, 3105±3108. Davis, M., Hut, P., Muller, R.A., 1984. Extinction of species by Grootes, P.M., Stulver, M., White, J.W.C., Johnsen, S., Jouzel, J., periodic comet showers. Nature 308, 715±717. 1993. Comparison of oxygen isotope records from the GISP2 Dessler, A.J., 1991. The small-comet hypothesis. Rev. Geophys. and GRIP Greenland ice cores. Nature 366, 552±554. 29, 355±382. Hallam, A., 1992. Phanerozoic Sea-level Changes. Columbia Dickens, G.R., Castillo, M.M., Walker, J.C.G., 1997. A blast of Univ. Press, New York, NY, 266 pp. gas in the latest Paleocene: simulating ®rst-order effects of Hannegan, B., Olsen, S., Prather, M., Zhu, X., Rind, D., Lerner, massive dissociation of oceanic methane hydrate. Geology 25, J., 1998. The dry stratosphere: a limit on cometary water 259±262. in¯ux. Geophys. Res. Lett. 25, 1649±1652. Donahue, T.M., 1986. Comment on the paper ªOn the in¯ux of Hays, J.D., Imbrie, J., Shackleton, N.J., 1976. Variations in the small comets into the Earth's upper atmosphere II. Interpreta- Earth's orbit: pacemaker of the ice ages. Science 194, 1121± tionº. Geophys. Res. Lett. 13, 555±557. 1132. Edwards, R.L., Cheng, H., Murrell, M.T., Goldstein, S.J., 1997. Hills, J.G., 1981. Comet showers and the steady-state infall of Protactinium-231 dating of carbonates by thermal ioniza- comets from the Oort cloud. Astron. J. 86, 1730±1740. tion mass spectrometry: implications for Quaternary climate Hsu, K.J., McKenzie, J.A., 1985. A `Strangelove' ocean in the change. Science 276, 782±786. earliest Tertiary. In: Sundquist, E.T., Broecker, W.S. (Eds.), Evans, D.A., Beukes, N.J., Kirschvink, J.L., 1997. Low-latitude The Carbon Cycle and Atmospheric CO2: Natural Variations glaciation in the Palaeoproterozoic era. Nature 386, 262±266. Archean to Present. Am. Geophys. Union Geophys. Monogr. Farley, K.A., Patterson, D.B., 1995. A 100-kyr periodicity in 32, 487±492. the ¯ux of extraterrestrial 3He to the sea ¯oor. Nature 378, Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, 600±603. A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., Fisher, D.M., 1996. Fabrics and veins in the forearc: a record 1984. The orbital theory of Pleistocene climate: support from of cyclic ¯uid ¯ow at depths of <15 km. In: Bebout, G.E. et a revised chronology of the marine δ18O record. In: Berger, al., (Eds.), Subduction Top to Bottom. Am. Geophys. Union A. et al. (Eds.), Milankovitch and Climate, Part 1. Reidel, Geophys. Monogr. 96, 75±89. Norwell, MA, pp. 269±305. Frakes, L.A., Francis, J.E., Syktus, J.I., 1992. Climate Modes of Imbrie, J., Boyle, E.A., Clemens, S.C., Duffy, A., Howard, W.R., the Phanerozoic. Cambridge Univ. Press, Cambridge, 274 pp. Kukla, G., Kutzbach, J., Martinson, D.G., McIntyre, A., Mix, Frank, L.A., 1990. The Big Splash. Carol Pub. Group, Seacau- A.C., Mol®no, B., Morley, J.J., Peterson, L.C., Pisias, N.G., cus, NJ, 255 pp. Prell, W.L., Raymo, M.E., Shackleton, N.J., Toggweiler, J.R., Frank, L.A., Sigwarth, J.B., 1993. Atmospheric holes and small 1992. On the structure and origin of major glaciation cycles 1. comets. Rev. Geophys. 31, 1±28. linear responses to Milankovitch forcing. Paleoceanography 7, Frank, L.A., Sigwarth, J.B., 1997a. Transient decreases of Earth's 701±738. far-ultraviolet dayglow. Geophys. Res. Lett. 24, 2423±2426. Imbrie, J., Berger, A., Boyle, E.A., Clemens, S.C., Duffy, A., Frank, L.A., Sigwarth, J.B., 1997b. Simultaneous observations of Howard, W.R., Kukla, G., Kutzbach, J., Martinson, D.G., transient decreases of Earth's far-ultraviolet dayglow with two McIntyre, A., Mix, A.C., Mol®no, B., Morley, J.J., Peterson, cameras. Geophys. Res. Lett. 24, 2427±2430. L.C., Pisias, N.G., Prell, W.L., Raymo, M.E., Shackleton, N.J., Frank, L.A., Sigwarth, J.B., 1997c. Detection of atomic oxygen Toggweiler, J.R., 1993a. On the structure and origin of major trails of small comets in the vicinity of Earth. Geophys. Res. glaciation cycles 2. the 100,000-year cycle. Paleoceanography Lett. 24, 2431±2434. 8, 699±735. Frank, L.A., Sigwarth, J.B., 1997d. Trails of OH emissions from Imbrie, J., Mix, A.C., Martinson, D.G., 1993b. Milankovitch small comets near Earth. Geophys. Res. Lett. 24, 2435±2438. theory viewed from Devil's Hole. Nature 363, 531±533. Frank, L.A., Sigwarth, J.B., 1998. Atmospheric holes: instrumen- Kennett, J.P., Stott, L.D., 1991. Abrupt deep-sea warming, tal and geophysical effects. J. Geophys. Res. (in press). palaeoceanographic changes and benthic extinctions at the Frank, L.A., Sigwarth, J.B., Craven, J.D., 1986. On the in¯ux of end of the Palaeocene. Nature 353, 225±229. small comets into the Earth's upper atmosphere II: interpreta- Knauth, L.P., Clemens, P.L., 1995. Climatic temperature history tion. Geophys. Res. Lett. 13, 307±310. of the Earth based on isotopic analyses of cherts. Geol. Soc. Frank, L.A., Sigwarth, J.B., Craven, J.D., 1987. Reply to ªCom- Am. Abstr. Progr. 27 (6), A205. ment on `On the In¯ux of Small Comets into the Earth's Upper Knauth, L.P., Lowe, D.R., 1978. Oxygen isotope geochemistry Atmosphere, II. Interpretation' by L.A. Frank, J.B. Sigwarth of cherts from the Onverwacht group (3.4 billion years), and J.D. Cravenº by J.T. Wasson and F.T. Kyte. Geophys. Res. Transvaal, South Africa, with implications for secular vara- Lett. 14, 781±782. tions in the isotopic composition of cherts. Earth Planet. Sci. Goodfriend, G.A., Mitterer, R.M., 1993. A 45,000-yr record Lett. 41, 209±222. of a tropical lowland biota: the land snail fauna from cave Kortenkamp, S.J., Dermott, S.F., 1998. A 100,000-year period- sediments at Coco Ree, Jamaica. Geol. Soc. Am. Bull. 105, icity in the accretion rate of interplanetary dust. Science 280, 18±29. 874±876. 50 D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51

Lebedinets, V.N., Kurbanmuradov, O., 1992. The role of Rampino, M.R., Haggerty, B.M., 1996. Impact crises and mass cometary and meteoritic matter in the genesis of noctilucent extinctions: a working hypothesis. In: Ryder, G. et al. (Eds.), clouds. Astron. Vestn. 26, 83±92, in Russian. The Cretaceous±Tertiary Event and Other Catastrophes in Lovelock, J.E., 1979. Gaia, a New Look at Life on Earth. Oxford Earth History. Geol. Soc. Am. Spec. Pap. 307, 11±30. Univ. Press, Oxford, 157 pp. Raymo, M.E., Ruddiman, W.F., 1992. Tectonic forcing of late Matese, J.J., Whitmire, D.P., 1996. Tidal imprint of distant galac- Cenozoic climate. Nature 359, 117±122. tic matter on the Oort comet cloud. Astrophys. J. Lett. 472, Raymo, M.E., Ruddiman, W.F., Froelich, P.N., 1988. In¯uence of L41±L43. late Cenozoic mountain building on ocean geochemical cycles. Meade, C., Jeanloz, R., 1991. Deep-focus earthquakes and re- Geology 16, 649±653. cycling of water into the Earth's mantle. Science 252, 68± Raynaud, D., Jouzel, J., Barnola, J.M., Chappellaz, J., Delmas, 72. R.J., Lorius, C., 1993. The ice record of greenhouse gasses. Milankovitch, M., 1930. Mathematische Klimalehre und As- Science 259, 926±934. tronomische Theorie der Klimaschwankungen. Borntraeger, Rizk, B., Dessler, A.J., 1997. Small comets: naked-eye visibility. Berlin, 176 pp. Geophys. Res. Lett. 24, 3121±3124. Miller, G.H., Magee, J.W., Jull, A.J.T., 1997. Low-latitude glacial Rubey, W.W., 1951. Geologic history of sea water, an attempt to cooling in the southern hemisphere from amino-acid racem- state the problem. Bull. Geol. Soc. Am. 62, 1111±1148. ization in emu eggshells. Nature 385, 241±244. Rye, R., Kuo, P.H., Holland, H.D., 1995. Atmospheric carbon Muller, R.A., MacDonald, G.J., 1995. Glacial cycles and orbital dioxide concentrations before 2.2 billions years ago. Nature inclinations. Nature 377, 107. 378, 603±605. Muller, R.A., MacDonald, G.J., 1997a. Simultaneous presence Schimel, D., 26 others, Radiative forcing of climate change. In: of orbital inclination and eccentricity in proxy climate records Houghton, J.T. et al. (Eds.), Climate Change 1995, The Sci- from Ocean Drilling Program Site 806. Geology 25, 3±6. ence of Climate Change. Cambridge Univ. Press, Cambridge, Muller, R.A., MacDonald, G.J., 1997b. Glacial cycles and astro- pp. 65±131. nomical forcing. Science 277, 215±218. Shaffer, J.A., Cerveny, R.S., Dorn, R.I., 1996. Radiation windows Muzio, L.J., Kramlich, J.C., 1988. An artifact in the measure- as indicators of an astronomical in¯uence on the Devil's Hole ment of N O from combustion sources. Geophys. Res. Lett. 2 chronology. Geology 24, 1017±1020. 15, 1369±1372. Slade, M.A., Butler, B.J., Muhleman, D.O., 1992. Mercury radar Nakamura, Y., Oberst, J., Clifford, S.M., Bills, B.G., 1986. imaging: evidence for polar ice. Science 258, 635±640. Comment on the letter ªOn the in¯ux of small comets into Smyth, J.R., 1994. A crystallographic model for hydrous wad- the Earth's upper atmosphereº. Geophys. Res. Lett. 13, 1184± sleyite (B-Mg SiO ): An ocean in the Earth's interior? Am. 1185. 2 4 Mineral. 79, 1021±1024. Nicholls, N., Gruza, G.V., Jouzel, J., Karl, T.R., Ogallo, L.A., Summers, M.E., Conway, R.R., Siskind, D.E., Stevens, M.H., Parker, D.E., 1996. Observed climate variability and change. In: Houghton, J.T. et al. (Eds.), Climate Change 1995, The Offermann, D., Riese, M., Preusse, D.F., Strobel III, J.M.R., Science of Climate Change. Cambridge Univ. Press, Cam- 1997. Implications of satellite OH observations for middle bridge, pp. 133±192. atmospheric H2O and Ozone. Science 277, 1967±1970. Nolet, G., Zielhuis, A., 1994. Low S velocities under the Sundquist, E.T., 1993. The global carbon dioxide budget. Science Tornquist±Teisseyre zone: evidence for water injection into the 259, 934±941. transition zone by subduction. J. Geophys. Res. 99, 15813± Swisher III, C.C., Grajales-Nishimura, J.M., Montanari, A., Mar- 15820. golis, S.V., Claeys, P., Alvarez, W., Renne, P., Cedillo-Pardo, Nozette, S., Lichtenberg, C.L., Spudis, P., Bonner, R., Ort, E., Maurrasse, F.J.-M.R., Curtis, G.H., Smit, J., McWilliams, 40 =30 W., Malaret, E., Robinson, M., Shoemaker, E.M., 1996. The M.O., 1992. Coeval Ar Ar ages of 65.0 million years Clementine bistatic radar experiment. Science 274, 1495± ago from Chicxulub crater melt rock and Cretaceous±Tertiary 1498. boundary tektites. Science 257, 954±958. Parks, G., Brittnacher, M., Chen, L.J., Elsen, R., McCarthy, M., Thomas, E., Shackleton, N.J., 1996. The Paleocene±Eocene ben- Germany, G., Spann, J., 1997. Does the UVI on Polar detect thic foraminiferal extinction and stable isotope anomalies. In: cosmic snowballs? Geophys. Res. Lett. 24, 3109±3112. Knox, R.W. O'B. et al. (Eds.), Correlation of the Early Pa- Parks, G., Brittnacher, M., Elsen, R., McCarthy, M., O'Meara, leogene in Northwest Europe. Geol. Soc. Spec. Publ. 101, J.M., Germany, G., Spann, J., 1998. Comparison of dark pixels 401±441. observed by VIS and UVI in dayglow images. Geophys. Res. Toon, O.B., Turco, R.P., Covey, C., 1997. Environmental pertur- Lett. 25, 3063±3066. bations caused by the impacts of asteroids and comets. Rev. Peucker-Ehrenbrink, B., 1996. Accretion of extraterrestrial mat- Geophys. 35, 41±78. ter during the last 80 million years and its effect on the Torgersen, T., Drenkard, S., Stute, M., Schlosser, P., Shapiro, marine osmium isotope record. Geochim. Cosmochim. Acta A., 1995. Mantle helium in ground waters of eastern North 60, 3187±3196. America: time and space constraints on sources. Geology 23, Rampino, M.R., 1994. Tillites, diamictites, and ballistic ejecta of 675±678. large impacts. J. Geol. 102, 439±456. Tuncel, G., Zoller, W.H., 1987. Atmospheric iridium at the South D. Deming / Palaeogeography, Palaeoclimatology, Palaeoecology 146 (1999) 33±51 51

Pole as a measure of the meteoritic component. Nature 329, Weissman, P.R., 1986. The Oort cloud and the galaxy: dynamical 703±705. interactions. In: Smoluchowski, R. et al. (Eds.), The Galaxy Turco, R.P., Toon, O.B., Park, C., Whitten, R.C., Pollack, J.B., and the Solar System. Univ. Arizona Press, Tucson, AZ, pp. 1982. An analysis of the physical, chemical, optical, and 204±237. historical impacts of the 1908 Tunguska Meteor Fall. Icarus Winograd, I.J., Coplen .B., T.B., Landwehr, J.M., Riggs, A.C., 50, 1±52. Ludwig, K.R., Szabo, B.J., Kolesar, P.T., Revesz, K.M., 1992. Vail, P.R., Mitchum, R.M., Jr., Thompson III, S., 1977. Seismic Continuous 500,00-year climate record from vein calcite in stratigraphy and global changes of sea level, Part 3: relative Devil's Hole, Nevada. Science 258, 255±260. changes of sea level from coastal onlap. In: Payton, C.E. Winograd, I.J., Landwehr, J.M., Ludwig, K.R., Coplen, T.B., (Ed.), Seismic Stratigraphy Ð Applications to Hydrocarbon Riggs, A.C., 1997. Duration and structure of the past four Exploration. Am. Assoc. Pet. Geol. Mem. 26, 83±97. interglaciations. Quat. Res. 48, 141±154.