What influence do astronomical phenomena have on the terrestrial climate and biosphere?

Coryn Bailer-JonesMax Planck Institute for Astronomy, Heidelberg

MPG LeadNet meeting, Berlin, 8 May 2012

Geological record: climate

variations on 10-100 Myr timescales

with evidence for widespread cooling and/or indirect orequivocal evidence for permanent ice represent cool Earthstates. An important distinction between the cool and coldEarth states is that the cool intervals have no unequivocalevidence for permanent ice. Many original literature sourc-es were incorporated in this analysis, but the compilationsof Frakes et al. (1992), Eyles (1993), Crowell (1999), Price(1999), and Isbell et al. (2003) proved most valuable. AllCO2 and climate records were calibrated to the timescaleof Gradstein et al. (2004); chronostratigraphic nomencla-ture also follows Gradstein et al. (2004). Formal statisticalanalyses are not presented here: a comparison betweenCO2 from proxies and geochemical models was presentedby Royer et al. (2004) but cannot be updated here owingto the uneven time-steps of the revised data (see Fig. 1 cap-tion). Statistical comparisons between CO2 and tempera-ture are also not possible because the temperature datapresented here are largely derived from regional studies,and so cannot be directly equated to global meantemperatures.

3. Results and discussion

3.1. Fidelity of Phanerozoic CO2 record

All 490 CO2 proxy records and their attendant errors areplotted in Fig. 1B, sorted by method; Fig. 1C plots the dataas a time series without error bars. When viewed at thescale of the Phanerozoic (Fig. 1B and C), the overarchingpattern is of high CO2 (4000+ ppm) during the early Paleo-zoic, a decline to present-day levels by the Pennsylvanian(!320 Ma), a rise to high values (1000–3000 ppm) duringthe Mesozoic, then a decline to the present-day. Thesebroad patterns are discernible even when the errors of indi-vidual data points (Fig. 1B) are taken into account.

When the CO2 proxy record is compared to the range ofreasonable CO2 predictions from the GEOCARB III mod-el (gray shaded region in Fig. 1C), it is clear that the vastmajority of proxy data fall within this uncertainty enve-lope. This provides support for the fidelity of the proxy re-cord. To explore this comparison further, Fig. 1D plots the‘best-guess’ predictions of GEOCARB III against a locallyweighted regression (LOESS) of the proxy data, where theLOESS is fitted to match the time-step of GEOCARB III.Again, a positive correlation is evident between these twoindependent records, and is consistent with previous anal-yses (Crowley and Berner, 2001; Royer et al., 2004). Wecan say with growing confidence that the broad, multi-mil-lion-year patterns of CO2 during much of the Phanerozoicare known.

If the GEOCARB and proxy CO2 records are used tocalculate radiative forcing (see Fig. 2 caption for details),the same general patterns remain: radiative forcing is thesame or weaker than pre-industrial conditions only duringthe two intervals of widespread, long-lived glaciation, thePermo-carboniferous and late Cenozoic (Fig. 2). The maindi!erence between the radiative forcing (Fig. 2) and CO2

(Fig. 1D) records is that radiative forcing is comparativelylow during the early Paleozoic owing to a weaker solar con-stant at that time.

3.2. Correlating CO2 to temperature: Late Ordovicianglaciation (Hirnantian; 445.6 – 443.7 Ma)

There is unequivocal evidence for a widespread but briefGondwanan glaciation during the end-Ordovician (Hir-nantian Stage; 445.6–443.7 Ma). Several reports argue fora longer interval of ice centered on the Ordovician–Silurianboundary (e.g., 58 my in Frakes et al., 1992), and alpineglaciers may have indeed persisted in Brazil and Boliviainto the early Silurian (Crowell, 1999), but most recentstudies demonstrate that the dominant glacial phase wasrestricted to the Hirnantian (Brenchley et al., 1994, 2003;Paris et al., 1995; Crowell, 1999; Sutcli!e et al., 2000).

There is one CO2 data point available that is close in ageto this glaciation, and it suggests very high CO2 levels(5600 ppm; see Fig. 3A; Yapp and Poths, 1992, 1996);moreover, GEOCARB III predicts high CO2 levels at thistime (!4200 ppm; see Fig. 1D). Apparently, this event pre-sents a critical test for the CO2-temperature paradigm (e.g.,Van Houten, 1985; Crowley and Baum, 1991). However, itis unclear what CO2 levels were during this event. Thesingle proxy record is Ashgillian in age, which spans theHirnantian but also most of the preceding Stage(450–443.7 Ma); if the CO2 data point dates to the pre-Hir-nantian Ashgillian, then this is consistent with a well-de-scribed mid-Ashgillian global warm event (Boucot et al.,2003; Fortey and Cocks, 2005). As for the insensitivity ofGEOCARB III to the glaciation, this is unsurprising giventhe brief duration of the event. Kump et al. (1999)

Time (Ma)0100200300400500

Rad

iativ

e fo

rcin

g(C

O2

+ s

un)

(W/m

2 )

-4

0

4

8 Proxies (LOESS)GEOCARB III

-8C O S D Carb P Tr J K Pg N

g

Fig. 2. Radiative forcing through the Phanerozoic. Radiative forcing isderived following the protocol of Crowley (2000a) and the radiativetransform expression for CO2 of Myhre et al. (1998). For the calculation,the CO2 records from Fig. 1D are used and solar luminosity is assumed tolinearly increase starting at 94.5% present-day values. Values are expressedrelative to pre-industrial conditions (CO2 = 280 ppm; solar luminosi-ty = 342 W/m2); a reference line of zero is given for clarity. The darkshaded bands correspond to periods with strong evidence for geograph-ically widespread ice (see Section 2 for details).

5668 D.L. Royer 70 (2006) 5665–5675

Royer (2006)

• 20-400 kyr periodicity (Milankovitch cycles)

‣ variation in eccentricity of Earth’s orbit

‣ also precession and variations in obliquity

Petit et al. (1999)

400 300 200 100 0

−8−6

−4−2

02

Time BP / kyr

Air t

empe

ratu

re v

aria

tion

/ deg

C

Geological record: impact cratering

0 500 1000 1500 2000

12

510

5020

0

Time before present / Myr

Dia

met

er /

km

170 km diameter,65 Myr old.

Mass extinctions?

Geological record: biodiversity

curve, and used species lists drawn from a large database offossil localities to standardize for sampling effort. The revisedPhanerozoic diversity curve seemed to them to be very differentfrom the Sepkoski curve. Jackson & Johnson (2001) raised athird issue concerning geographical bias and the poor representa-tion of high-diversity, low-latitude marine faunas in currentdatabases. Their recent collections from rocks in a small part ofthe Caribbean had shown unexpectedly high levels of diversityamongst Plio-Pleistocene invertebrates. Because the best-described regional faunas of this time interval lay outside thetropics, Jackson & Johnson (2001) concluded that new data wereneeded to overcome geographical biases and obtain a trueindication of how marine diversity had changed. Finally, Smith etal. (2001) provided an empirical case study showing how biasesin habitat representation in the rock record could seriously distortglobal marine diversity curves. Since then, a steady stream ofpapers have appeared questioning and probing Phanerozoicmarine diversity patterns.In this review I shall briefly explain the problems that face

palaeontologists wishing to estimate marine diversity throughtime, review some of the techniques currently being developed toovercome these problems, and end by looking at a couple ofaspects of the Phanerozoic marine diversity curve that are nowunder intense scrutiny.

What is wrong with the way marine diversity has beenestimated in the past?

Prior to 2001 Phanerozoic diversity curves were constructed froma simple count of numbers of taxa recorded in any given timeinterval (usually the 72–77 stage-level intervals of Sepkoski

(1982) and Benton (1993)). Compilations at any taxonomic levelcan be used to construct diversity curves, but Robeck et al. (2000)demonstrated that using more finely subdivided taxonomic group-ings produced a more precise view of underlying diversity in therocks. Furthermore, although there will always be a certainamount of error in taxonomic compilations, Sepkoski (1993) andAdrain & Westrop (2000) both demonstrated that such error wasrandom and thus did not pose a serious problem to this approach.

The fossil record is of course notoriously incomplete, so tocompensate for this a technique called range interpolation hasbeen employed. Range interpolation removes some problems of apatchy fossil record by assuming that a taxon is present in eachtime interval between its first and last occurrence, whether or notit has actually been found in those time intervals. Because thefossil record is dominated by organisms with mineralizedskeletons, the history of those taxa with hard parts is taken as aproxy for all marine diversity. The exact ratio of mineralized tounmineralized taxa is unimportant so long as it has remainedbroadly similar throughout the Phanerozoic. By assuming thatsampling is more or less uniform through time, the relativenumbers of taxa described from each time interval (or that crossboundaries between intervals) can be used as a measure of howdiversity has changed.

This taxon-counting approach is simple to employ and see-mingly robust to certain potential problems, but makes thefollowing three critical assumptions: (1) all time intervals areequally well sampled; (2) preservation potential is uniform overtime; (3) taxonomists partition taxa in a uniform manner. Eachunfortunately is beset with problems.

Sampling of the rock record

There are two aspects of sampling that need to be considered:geographical bias and variation in sampling intensity.

Geographical bias. Jackson & Johnson (2001) and Johnson(2003) argued that any diversity curve constructed simply fromcataloguing the numbers of fossils already described was doomedto failure because well-studied parts of the world that contributemost to taxonomic compilations were not necessarily representa-tive of global diversity. Specifically, they found that the Neogenerecord of the tropics was woefully undersampled compared withtemperate regions, a view later reinforced by Valentine et al.(2006).

European and North American data certainly contribute dis-proportionately to taxonomic compilations, simply because fossilcollecting has been intensely pursued in those regions for muchlonger (Kidwell & Holland 2002). However, extreme unbalancein sampling between, say, Indo-Pacific faunas and those oftemperate North America is no problem if this bias appliesequally to all time intervals through the Phanerozoic.

Unfortunately, continental plates have migrated out of thetropics over time (Allison & Briggs 1993; Walker et al. 2002;Fig. 2). Because diversity is highest in the tropics, a long-termtrend of decreasing diversity could be created artificially simplybecause the well-studied parts of the world have shifted over timefrom equatorial to temperate latitude through plate migration.Indeed, some palaeontologists are starting to factor out this biasfrom their analyses (e.g. Bush & Bambach 2004; Crampton et al.2006b). On the other hand, the smaller-scale rises and falls indiversity from stage to stage that have been taken as thesignatures of mass extinction and radiation cannot be explainedby such slow changes in the positions of continental blocks(Smith 2001).

Fig. 1. Phanerozoic diversity curves derived from counting the number

of taxa present in each stage, with range interpolation. (a) Genus-level

diversity, from Sepkoski (2002). (b) Family-level diversity from Benton

(1995) and Sepkoski (1997).

A. B. SMITH732

Smith (2007)

today550 Myr BP

Example claim of periodicity in geological time series

• biodiversity (Rohde & Muller 2005)

• significant period of 62 ± 3 Myr claimed

9. Newman, A. V. et al. Along-strike variability in the seismogenic zone below Nicoya Peninsula, Costa

Rica. Geophys. Res. Lett. 29, 38–41 (2002).

10. Chadwell, C. D. & Bock, Y. Direct estimation of absolute precipitable water in oceanic regions by GPS

tracking of a coastal buoy. Geophys. Res. Lett. 28, 3701–3704 (2001).

11. Spiess, F. N. et al. Precise GPS/acoustic positioning of seafloor reference points for tectonic studies.

Physics Earth Planet. Inter. 108, 101–112 (1998).

12. Webb, F. H. & Zumberge, J. F. An introduction to GIPSY/OASIS-II (JPL Publication D-11088, Jet

Propulsion Lab., Pasadena, California, 1997).

13. Chadwell, C. D. Shipboard towers for Global Positioning System antennas.Ocean Eng. 30, 1467–1487

(2003).

14. Gomberg, J. & Ellis, M. Topography and tectonics of the central NewMadrid seismic zone: Results of

numerical experiments using a three-dimensional boundary-element program. J. Geophys. Res. 99,

20299–20310 (1994).

15. Krabbenhoft, A., Bialas, J., Kopp, H., Kukowski, N. & Hubscher, C. Crustal structure of the Peruvian

continental margin from wide-angle seismic studies. Geophys. J. Int. 159, 749–764 (2004).

16. Hampel, A., Kukowski, N., Bialas, J., Huebscher, C. & Heinbockel, R. Ridge subduction at an erosive

margin: The collision of the Nazca Ridge in southern Peru. J. Geophys. Res. 109, B02101, doi:10.1029/

2003JB002593 (2004).

17. Sella, G., Dixon, T. & Mao, A. REVEL: A model for recent plate velocites from space geodesy.

J. Geophys. Res. 107, 2081, doi:10.1029/2000JB00033 (2002).

18. Angermann, D. & Klotz, J. R. Space geodetic estimation of the Nazca–South America Euler vector.

Earth Planet. Sci. Lett. 171, 329–334 (1999).

19. DeMets, C., Gordon, R., Argus, D. & Stein, S. Effect of recent revision to the geomagnetic reversal time

scale on estimates of current plate motion. Geophys. Res. Lett. 21, 2191–2194 (1994).

20. Larson, K. M., Freymueller, J. T. & Philipsen, S. Global plate velocities from the Global Positioning

System. J. Geophys. Res. 102, 9961–9981 (1997).

21. Wang, K. & Dixon, T. “Coupling” semantics and science in earthquake research. Eos 85, 180

(2004).

22. Tichelaar, B. & Ruff, L. Seismic coupling along the Chilean subduction zone. J. Geophys. Res. 96,

11997–12022 (1991).

23. Bevis, M., Smalley, R. Jr, Herring, T., Godoy, J. & Galban, F. Crustal motion north and south of the

Arica Deflection: Comparing recent geodetic results from the Central Andes. Geochem. Geophys.

Geosyst. 1, 1999GC000011 (1999).

24. Schweller,W. J., Kulm, L. D. & Prince, R. A. inNazca Plate: Crustal Formation and Andean Convergence

(eds Kulm, L. D., Dymond, J., Dasch, E. J., Hussong, D. M. & Roderick, R.) 323–349 (Mem. Geol. Soc.

Am. 154, Geological Society of America, Boulder, Colorado, 1981).

25. Altimini, A., Sillard, P. & Boucher, C. ITRF2000: A new release of the International Terrestrial

Reference Frame for earth science applications. J. Geophys. Res. 107, 2214, doi:10.1029.2001JB000561

(2002).

26. Smith, W. H. F. & Sandwell, D. T. Global seafloor topography from satellite altimetry and ship depth

soundings. Science 277, 1957–1962 (1997).

Supplementary Information accompanies the paper on www.nature.com/nature.

Acknowledgements We thank M. Bevis for comments and suggestions; R. Zimmerman,D. Rimington and D. Price for engineering support; and the Captain and crew of the R/V RogerRevelle. We thank the Instituto Geofisico Del Peru for operating the land GPS stations and theInstituto Del Mar Del Peru, Direccion de Higrografia y Navagacion, for support at sea. This workwas supported by the Marine Geology and Geophysics Program of the US National ScienceFoundation.

Competing interests statement The authors declare that they have no competing financialinterests.

Correspondence and requests for materials should be addressed to C.D.C. (cchadwell@ucsd.edu).

..............................................................

Cycles in fossil diversityRobert A. Rohde & Richard A. Muller

Department of Physics and Lawrence Berkeley Laboratory, University ofCalifornia, Berkeley, California 94720, USA.............................................................................................................................................................................

It is well known that the diversity of life appears to fluctuateduring the course of the Phanerozoic, the eon during which hardshells and skeletons left abundant fossils (0–542 million yearsago). Here we show, using Sepkoski’s compendium1 of the firstand last stratigraphic appearances of 36,380 marine genera, astrong 62 6 3-million-year cycle, which is particularly evident inthe shorter-lived genera. The five great extinctions enumeratedby Raup and Sepkoski2 may be an aspect of this cycle. Because ofthe high statistical significance we also consider the contri-butions of environmental factors, and possible causes.

Sepkoski’s posthumously published Compendium of FossilMarine Animal Genera1, and its earlier versions, has frequentlybeen used in the study of biodiversity and extinction3,4. For ourpurposes, diversity is defined as the number of distinct genera aliveat any given time; that is, those whose first occurrence predatesand whose last occurrence postdates that time. Because Sepkoskireferences only 295 stratigraphic intervals, the International Com-mission on Stratigraphy’s 2004 time scale5 is used to translate thestratigraphic references into a record of diversity versus time; detailsare given in the Supplementary Information. Although Sepkoski’s isthe most extensive compilation available, it is known to be subjectto certain systematic limitations due primarily to the varyingavailability and quality of geological sections6,7. The implicationsof this will be discussed where appropriate.

Figure 1a shows a plot of diversity against time for all 36,380genera in Sepkoski’s Compendium. In Fig. 1b we show the 17,797genera that remain when we remove those with uncertain ages(given only at epoch or period level), and those with only a singleoccurrence. The smooth trend curve through the data is the third-order polynomial that minimizes the variance of the difference

Figure 1 Genus diversity. a, The green plot shows the number of known marine animalgenera versus time from Sepkoski’s compendium1, converted to the 2004 Geologic Time

Scale5. b, The black plot shows the same data, with single occurrence and poorly datedgenera removed. The trend line (blue) is a third-order polynomial fitted to the data. c, As b,with the trend subtracted and a 62-Myr sine wave superimposed. d, The detrended dataafter subtraction of the 62-Myr cycle and with a 140-Myr sine wave superimposed.

Dashed vertical lines indicate the times of the five major extinctions2. e, Fourier spectrumof c. Curves W (in blue) and R (in red) are estimates of spectral background. Conventional

symbols for major stratigraphic periods are shown at the bottom.

letters to nature

NATURE | VOL 434 | 10 MARCH 2005 | www.nature.com/nature208©!!""#!Nature Publishing Group!

!

Age / Myr0 550

Num

ber

of g

ener

a

Suggested astronomical mechanisms

Perturbations of Oortcloud by Galactic tideand/or passing stars⇒ comet impacts

Nearby supernovae⇒ gamma rays

⇒ biological extinction

Star forming regions⇒ cosmic rays⇒ cloud formation(highly questionable!)

Suggested causes of periodicity

• motion of the Sun in the Galaxy

‣ vertical oscillation through disk (periods of 50-75 Myr)

‣ spiral arm crossing (timescale of 50-100 Myr)

picture credit: Medvedev

Diamonds along the Sun’s track indicate its placement at inter-vals of 100 Myr. We see that for this assumed pattern speed, theSun has passed through only two arms over the last 500 Myr.However, if we assume a lower but still acceptable pattern speedof !p ! 14:4 km s"1 kpc"1 (shown in Fig. 3 for !# " !p !11:9 km s"1 kpc"1), then the Sun has crossed four spiral arms inthe past 500 Myr and has nearly completed a full rotation aheadof the spiral pattern. Thus, the choice of the spiral pattern speeddramatically influences any conclusions about the number andtiming of the Sun’s passages through the spiral arms over thistime interval.

The duration of a coherent spiral pattern is an open question,but there is some evidence that long-lived spiral patterns may bemore prevalent in galaxies with a central bar. For example, numer-ical simulations of the evolution of barred spirals by Rautiainen&Salo (1999) suggest that spiral patternsmay last several gigayears.Their work suggests that the shortest timescale for the appearanceor disappearance of a spiral arm is about 1 Gyr. Therefore, it is rea-sonable to assume that the present day spiral structure has prob-ably been more or less intact over the last 500 Myr (at least in theregion of the solar circle).

3. DISCUSSION

Shaviv (2003) argues that the Earth has experienced fourlarge-scale cycles in the CRF over the last 500 Myr (with sim-

ilar cycle times back to 1 Gyr before the present). Shaviv showsthat the CRF exposure ages of iron meteorites indicate a peri-odicity of 143 $ 10 Myr in the CRF rate. Since the cosmic-rayproduction is related to supernovae, and since Type II super-novae will be more prevalent in the young star-forming regionsof the spiral arms, Shaviv suggests that the periodicity corre-sponds to the mean time between arm crossings (so that Earthhas made four arm crossings over the last 500 Myr). Shaviv(2003) and Shaviv & Veizer (2003) show how the epochs ofenhanced CRF are associated with cold periods on Earth. Thegeological record of climate-sensitive sedimentary layers (gla-cial deposits) and the paleolatitudinal distribution of ice-rafteddebris (Frakes et al. 1992; Crowell 1999) indicate that the Earthhas experienced periods of extended cold (‘‘icehouses’’) and hottemperatures (‘‘greenhouses’’) lasting tens of millions of years(Frakes et al. 1992). The long periods of cold may be punctuatedby much more rapid episodes of ice age advances and declines(Imbrie et al. 1992). The climate variations indicated by the geo-logical evidence of glaciation are confirmed by measurements ofancient tropical sea temperatures, through oxygen isotope lev-els in biochemical sediments (Veizer et al. 2000). All of thesestudies lead to a generally coherent picture in which four peri-ods of extended cold have occurred over the last 500 Myr, andthe midpoints of these ice age epochs (IAEs) are summarizedin Table 1 (see Shaviv 2003). The icehouse times according toFrakes et al. (1992) are indicated by the thick solid line segmentsin Figures 1–3.If these IAEs do correspond to the Sun’s passages through

spiral arms, then it is worthwhile considering which spiral pat-tern speeds lead to crossing times during ice ages. We calcu-lated the crossing times for a grid of assumed values of!#" !p

and found the value that minimized the !2" residuals of the dif-

ferences between the crossing times and IAEs. There are twomajor error sources in the estimation of the timing differences.First, the calculated arm crossing times depend sensitively on theplacement of the spiral arms, and we made a comparison betweenthe crossing times for our adopted model and that of Russeil(2003) to estimate the timing error related to uncertainties in theposition of the spiral arms (approximately$8 Myr except in thecase of the crossing of the Scutum-Crux arm on the far side ofthe Galaxy, where the difference is%40Myr). Secondly, there areerrors associated with the estimated midtimes of the IAEs, and weused the scatter between the various estimates in columns (2)–(5)of Table 1 to set this error (approximately$14Myr). We adoptedthe quadratic sum of these two errors in evaluating the!2

" statisticof each fit. The results of the fitting procedure for various modeland sample assumptions are listed in Table 2.The first trial fit was made by finding the !2

" minimum thatbest matched the crossing times with the IAE midpoints fromShaviv (2003; given in col. [5] of Table 1 and noted as ‘‘Mid-point’’ in col. [2] of Table 2). All four arm crossings wereincluded in the calculation (indicated as 1–4 in col. [3] of

Fig. 3.—Depiction of the Sun’s motion relative to the spiral arm pattern,in the same format as Fig. 2 but for a smaller spiral pattern speed (!p !14:4 km s"1 kpc"1).

TABLE 1

Midpoints of Ice Age Epochs

Ice Age Epoch

(1)

Crowell (1999)(Myr BP)

(2)

Frakes et al. (1992)(Myr BP)

(3)

Veizer et al. (2000)(Myr BP)

(4)

Shaviv (2003)(Myr BP)

(5)

Arm Crossing (Fit 2)(Myr BP)

(6)

1.................................. <22 <28 30 20 80

2.................................. 155 144 180 160 156

3.................................. 319 293 310 310 3104.................................. 437 440 450 446 446

GIES & HELSEL846 Vol. 626

Gies & Helsel (2005)

observedgeologicaltime series

predictedgeologicaltime series

Bayesianmodel

assessment

astronomicalobservations

Galaxymodel

mechanisms:supernovae, impacts,

cosmic rays, etc.

solar orbitand solar

environment

Results:evidencefor model

alternative models:periodic, trend,

noise, “internal”, etc.

astroimpacts research programme at MPIA

parametrizedmodel

Coryn Bailer-Jones, Max Planck Institute for Astronomy, Heidelberg

Summary, and conclusions so far

• there are plausible astronomical mechanisms to influence terrestrial climate and biosphere, but

• many claims based on poor statistical analyses (CBJ 2009)

• terrestrial impact cratering (CBJ 2011)

‣ use of more rigorous (Bayesian) methods of model comparison

‣ refutes claims of periodicity in cratering

• ongoing project to model geological time series

‣ testing of astronomical, terrestrial and noise models

• more information and references: www.astroimpacts.org