The Silurian Hypothesis: Would it be possible to detect an industrial civilization in the geological record?
((in press Int. J. Astrobio)doi: 10.1017/S1473550418000095.
The Silurian Hypothesis: Would it be possible to detectan industrial civilization in the geological record?
Gavin A. Schmidt and Adam Frank NASA Goddard Institute for Space Studies, 2880 Broadway, New York, NY 10025 Department of Physics and Astronomy, University of Rochester, Rochester NY 14620
Abstract
If an industrial civilization had existed on Earth many millions of years prior to ourown era, what traces would it have left and would they be detectable today? Wesummarize the likely geological fingerprint of the Anthropocene, and demonstratethat while clear, it will not differ greatly in many respects from other known eventsin the geological record. We then propose tests that could plausibly distinguish anindustrial cause from an otherwise naturally occurring climate event.
Keywords:
Astrobiology – Drake Equation – industrial civilization – Silurian hypothesis– Anthropocene – PETM a r X i v : . [ a s t r o - ph . E P ] A p r he Silurian Hypothesis The search for life elsewhere in the universe is a central occupation of astrobiology and scientistshave often looked to Earth analogues for extremophile bacteria, life under varying climate statesand the genesis of life itself. A subset of this search is the prospect for intelligent life, and then afurther subset is the search for civilizations that have the potential to communicate with us. Acommon assumption is that any such civilization must have developed industry of some sort. Inparticular the ability to harness those industrial processes to develop radio technologies capable ofsending or receiving messages. In what follows, however, we will define industrial civilizations hereas the ability to harness external energy sources at global scales.One of the key questions in assessing the likelihood of finding such a civilization is an understand-ing of how often, given that life has arisen and that some species are intelligent, does an industrialcivilization develop? Humans are the only example we know of, and our industrial civilization haslasted (so far) roughly 300 years (since, for example, the beginning of mass production methods).This is a small fraction of the time we have existed as a species, and a tiny fraction of the timethat complex life has existed on the Earth’s land surface ( ∼
400 million years ago, Ma). This shorttime period raises the obvious question as to whether this could have happened before. We termthis the "Silurian Hypothesis" .While much idle speculation and late night chatter has been devoted to this question, we areunaware of previous serious treatments of the problem of detectability of prior terrestrial industrialcivilizations in the geologic past. Given the vast increase in work surrounding exoplanets andquestions related to detection of life, it is worth addressing the question more formally and in itsastrobiological setting. We note also the recent work of Wright (2017) which addressed aspectsof the problem and previous attempts to assess the likelihood of solar system non-terrestrialcivilization such as Haqq-Misra & Kopparapu (2012). This paper is an attempt to remedy thegap in a way that also puts our current impact on the planet into a broader perspective. We first We name the hypothesis after a 1970 episode of the British science fiction TV series Doctor Who where a longburied race of intelligent reptiles "Silurians" are awakened by an experimental nuclear reactor. We are not howeversuggesting that intelligent reptiles actually existed in the Silurian age, nor that experimental nuclear physics isliable to wake them from hibernation. Other authors have dealt with this possibility in various incarnations (forinstance, Hogan (1977)), but it is a rarer theme than we initially assumed. he Silurian Hypothesis
The Drake equation is the well-known framework for estimating of the number of active, commu-nicative extraterrestrial civilizations in the Milky Way galaxy (Drake, 1961, 1965). The number ofsuch civilizations, N, is assumed to be equal to the product of; the average rate of star formation,R ∗ , in our galaxy; the fraction of formed stars, f p , that have planets; the average number ofplanets per star, n e , that can potentially support life; the fraction of those planets, f l , that actuallydevelop life; the fraction of planets bearing life on which intelligent, civilized life, f i , has developed;the fraction of these civilizations that have developed communications, f c , i.e., technologies thatrelease detectable signs into space, and the length of time, L , over which such civilizations releasedetectable signals. N = R ∗ · f p · n e · f ‘ · f i · f c · L If over the course of a planet’s existence, multiple industrial civilizations can arise over thespan of time that life exists at all, the value of f c may in fact be greater than one.This is a particularly cogent issue in light of recent developments in astrobiology in whichthe first three terms, which all involve purely astronomical observations, have now been fullydetermined. It is now apparent that most stars harbor families of planets (Seager, 2013). Indeed,many of those planets will be in the star’s habitable zones (Howard, 2013; Dressing & Charbonneau,2013). These results allow the next three terms to be bracketed in a way that uses the exoplanetdata to establish a constraint on exo-civilization pessimism. In Frank & Sullivan (2016) such a“pessimism line” was defined as the maximum "biotechnological" probability (per habitable zoneplanets) f bt for humans to be the only time a technological civilization has evolved in cosmic history.Frank & Sullivan (2016) found f bt in the range ∼ − to 10 − . he Silurian Hypothesis f ‘ , f i and f c . Earth’s history often serves as a template for discussions of possible values for theseprobabilities. For example the has been considerable discussion of how many times life began onEarth during the early Archean given the ease of abiogenisis (Patel et al., 2015) including thepossibility of a "shadow biosphere" composed of descendants of a different origin event from theone which led to our Last Universal Common Ancestor (LUCA) (Cleland & Copley, 2006). Inaddition, there is a long standing debate concerning the number of times intelligence has evolvedin terms of dolphins and other species (Marino, 2015). Thus only the term f c has been commonlyaccepted to have a value on Earth of strictly 1. Consideration of previous civilizations on other solar system worlds has been taken on by Wright(2017) and Haqq-Misra & Kopparapu (2012). We note here that abundant evidence exists ofsurface water in ancient Martian climates (3.8 Ga) (e.g. Achille & Hynek, 2010; Arvidson et al.,2014), and speculation that early Venus (2 Ga to 0.7 Ga) was habitable (due to a dimmer sunand lower CO atmosphere) has been supported by recent modeling studies (Way et al., 2016).Conceivably, deep drilling operations could be carried out on either planet in future to assess theirgeological history. This would constrain consideration of what the fingerprint might be of life, andeven organized civilization (Haqq-Misra & Kopparapu, 2012). Assessments of prior Earth eventsand consideration of Anthropocene markers such as those we carry out below will likely provide akey context for those explorations. That this paper’s title question is worth posing is a function of the incompleteness of the geologicalrecord. For the Quaternary (the last 2.5 million years), there is widespread extant physical evidenceof, for instance, climate changes, soil horizons, and archaeological evidence of non-Homo Sapienscultures (Denisovians, Neanderthals etc.) with occasional evidence of bipedal hominids datingback to at least 3.7 Ma (e.g. the Laetoli footprints) (Leakey & Hay, 1979). The oldest extant he Silurian Hypothesis ∼
170 Ma) (ODP Leg 801 Team,2000).The fraction of life that gets fossilized is always extremely small and varies widely as afunction of time, habitat and degree of soft tissue versus hard shells or bones (Behrensmeyer et al.,2000). Fossilization rates are very low in tropical, forested environments, but are higher in aridenvironments and fluvial systems. As an example, for all the dinosaurs that ever lived, there areonly a few thousand near- complete specimens, or equivalently only a handful of individual animalsacross thousands of taxa per 100,000 years. Given the rate of new discovery of taxa of this age,it is clear that species as short-lived as Homo Sapiens (so far) might not be represented in theexisting fossil record at all.The likelihood of objects surviving and being discovered is similarly unlikely. Zalasiewicz (2009)speculates about preservation of objects or their forms, but the current area of urbanization isless than 1% of the Earth’s surface (Schneider et al., 2009), and exposed sections and drillingsites for pre-Quaternary surfaces are orders of magnitude less as fractions of the original surface.Note that even for early human technology, complex objects are very rarely found. For instance,the Antikythera Mechanism (ca. 205 BCE) is a unique object until the Renaissance. Despiteimpressive recent gains in the ability to detect the wider impacts of civilization on landscapesand ecosystems (Kidwell, 2015), we conclude that for potential civilizations older than about 4Ma, the chances of finding direct evidence of their existence via objects or fossilized examples oftheir population is small. We note, however, that one might ask the indirect question related toantecedents in the fossil record indicating species that might lead downstream to the evolution oflater civilization-building species. Such arguments, for or against, the Silurian hypothesis wouldrest on evidence concerning highly social behavior or high intelligence based on brain size. Theclaim would then be that there are other species in the fossil record which could, or could not, haveevolved into civilization-builders. In this paper, however, we focus on physico-chemical tracers for he Silurian Hypothesis
We will restrict the scope of this paper to geochemical constraints on the existence of pre-Quaternaryindustrial civilizations, that may have existed since the rise of complex life on land. This rules outsocieties that might have been highly organized and potentially sophisticated but that did notdevelop industry and probably any purely ocean-based lifeforms. The focus is thus on the periodbetween the emergence of complex life on land in the Devonian ( ∼
400 Ma) in the Paleozoic eraand the mid-Pliocene ( ∼ While an official declaration of the Anthropocene as a unique geological era is still pending (Crutzen,2002; Zalasiewicz et al., 2017), it is already clear that our human efforts will impact the geologicrecord being laid down today (Waters et al., 2014). Some of the discussion of the specific boundarythat will define this new period is not relevant for our purposes because the markers proposed(ice core gas concentrations, short-half-lived radioactivity, the Columbian exchange) (e.g. Lewis& Maslin, 2015; Hamilton, 2016) are not going to be geologically stable or distinguishable onmulti-million year timescales. However, there are multiple changes that have already occurred thatwill persist. We discuss a number of these below.There is an interesting paradox in considering the Anthropogenic footprint on a geologicaltimescale. The longer human civilization lasts, the larger the signal one would expect in therecord. However, the longer a civilization lasts, the more sustainable its practices would needto have become in order to survive. The more sustainable a society (e.g. in energy generation,manufacturing, or agriculture) the smaller the footprint on the rest of the planet. But the smallerthe footprint, the less of a signal will be embedded in the geological record. Thus the footprint ofcivilization might be self-limiting on a relatively short time-scale. To avoid speculating about theultimate fate of humanity, we will consider impacts that are already clear, or that are foreseeable he Silurian Hypothesis
Since the mid-18th Century, humans have released over 0.5 trillion tons of fossil carbon via theburning of coal, oil and natural gas (Le Quéré et al., 2016), at a rate orders of magnitude fasterthan natural long-term sources or sinks. In addition, there has been widespread deforestation andaddition of carbon dioxide into the air via biomass burning. All of this carbon is biological inorigin and is thus depleted in C compared to the much larger pool of inorganic carbon (Revelle& Suess, 1957). Thus the ratio of C to C in the atmosphere, ocean and soils is decreasing (animpact known as the “Suess Effect” (Quay et al., 1992)) with a current change of around -1 (cid:104) δ Csince the pre-industrial (Böhm et al., 2002; Eide et al., 2017) in the surface ocean and atmosphere(figure 1a).As a function of the increase of fossil carbon into the system, augmented by black carbonchanges, other non-CO trace greenhouse gases (like N O, CH and chloro-fluoro-carbons (CFCs)),global industrialization has been accompanied by a warming of about 1 ◦ C so far since the mid19th Century (GISTEMP Team, 2016; Bindoff et al., 2013). Due to the temperature-relatedfractionation in the formation of carbonates (Kim & O’Neil, 1997) (-0.2 (cid:104) δ O per ◦ C) and strongcorrelation in the extra-tropics between temperature and δ O (between 0.4 and 0.7 (cid:104) per ◦ C)(and roughly 8 × as sensitive for deuterium isotopes relative to hydrogen ( δ D)), we expect thistemperature rise to be detectable in surface ocean carbonates (notably foraminifera), organicbiomarkers, cave records (stalactites), lake ostracods and high-latitude ice cores, though only thefirst two of these will be retrievable in the time-scales considered here.The combustion of fossil fuel, the invention of the Haber-Bosch process, the large-scale appli-cation of nitrogenous fertilizers, and the enhanced nitrogen fixation associated with cultivated he Silurian Hypothesis δ Nanomalies are already detectable in sediments remote from civilization (Holtgrieve et al., 2011).
There are multiple causes of a greatly increased sediment flow in rivers and therefore in depositionin coastal environments. The advent of agriculture and associated deforestation have lead to largeincreases in soil erosion (Goudie, 2000; National Research Council, 2010). Furthermore, canalizationof rivers (such as the Mississippi) have led to much greater oceanic deposition of sediment thanwould otherwise have occurred. This tendency is mitigated somewhat by concurrent increases inriver dams which reduce sediment flow downstream. Additionally, increasing temperatures andatmospheric water vapor content have led to greater intensity of precipitation (Kunkel et al., 2013)which, on its own, would also lead to greater erosion, at least regionally. Coastal erosion is also onthe increase as a function of rising sea level, and in polar regions is being enhanced by reductionsin sea ice and thawing permafrost (Overeem et al., 2011).In addition to changes in the flux of sediment from land to ocean, the composition of thesediment will also change. Due to the increased dissolution of CO in the ocean as a functionof anthropogenic CO emissions, the upper ocean is acidifying (a 26% increase in H + or 0.1 pHdecrease since the 19th Century) (Orr et al., 2005). This will lead to an increase in CaCO dissolution within the sediment that will last until the ocean can neutralize the increase. There willalso be important changes in mineralogy (Zalasiewicz et al., 2013; Hazen et al., 2017). Increases incontinental weathering are also likely to change ratios of strontium and osmium (e.g. Sr/ Srand
Os/
Os ratios) (Jenkyns, 2010).As discussed above, nitrogen load in rivers is increasing as a function of agricultural practices.This in turn is leading to more microbial activity in the coastal ocean which can deplete dissolvedoxygen in the water column (Diaz & Rosenberg, 2008), and recent syntheses suggests a globaldecline already of about 2% (Schmidtko et al., 2017; Ito et al., 2017). This in turn is leading toan expansion of the oxygen minimum zones, greater ocean anoxia, and the creation of so-called“dead-zones” (Breitburg et al., 2018). Sediment within these areas will thus have greater organic he Silurian Hypothesis
The last few centuries have seen significant changes in the abundance and spread of small animals,particularly rats, mice, and cats etc., that are associated with human exploration and bioticexchanges. Isolated populations almost everywhere have now been superseded in many respectsby these invasive species. The fossil record will likely indicate a large faunal radiation of theseindicator species at this point. Concurrently, many other species have already, or are likely tobecome, extinct, and their disappearance from the fossil record will be noticeable. Given theperspective from many million years ahead, large mammal extinctions that occurred at the end ofthe last ice age will also associated with the onset of the Anthropocene.
There are many chemicals that have been (or were) manufactured industrially that for variousreasons can spread and persist in the environment for a long time (Bernhardt et al., 2017). Mostnotably, persistent organic pollutants (POPs) (organic molecules that are resistant to degradationby chemical, photo-chemical or biological processes), are known to have spread across the world(even to otherwise pristine environments) (Beyer et al., 2000). Their persistence is often tied tobeing halogenated organics since the bond strength of C-Cl (for instance) is much stronger thanC-C. For instance, polychlorinated biphenyls (PCBs) are known to have lifetimes of many hundredsof years in river sediment (Bopp, 1979). How long a detectable signal would persist in oceansediment is, however, unclear.Other chlorinated compounds may also have the potential for long-term preservation, specificallyCFCs and related compounds. While there are natural sources for the most stable compound he Silurian Hypothesis ), there are only anthropogenic sources for C F and SF , the next most stable compounds.In the atmosphere, their sink via photolytic destruction in the stratosphere limits their lifetimes toa few thousand years (Ravishankara et al., 1993). The compounds do dissolve in the the oceanand can be used as tracers of ocean circulation, but we are unaware of studies indicating how longthese chemicals might survive and/or be detectable in ocean sediment given some limited evidencefor microbial degradation in anaerobic environments (Denovan & Strand, 1992).Other classes of synthetic biomarkers may also persist in sediments. For instance, steroids, leafwaxes, alkenones, and lipids can be preserved in sediment for many millions of years (i.e Paganiet al., 2006). What might distinguish naturally occurring biomarkers from synthetics might be thechirality of the molecules. Most total synthesis pathways do not discriminate between D- and L-chirality, while biological processes are almost exclusively monochiral (Meierhenrich, 2008) (forinstance, naturally occurring amino acids are all L-forms, and almost all sugars are D-forms).Synthetic steroids that do not have natural counterparts are also now ubiquitous in water bodies. Since 1950, there has been a huge increase in plastics being delivered into the ocean (Moore,2008; Eriksen et al., 2014). Although many common forms of plastic (such as polyethylene andpolypropylene) are buoyant in sea water, and even those that are nominally heavier than watermay be incorporated into flotsam that remains at the surface, it is already clear that mechanicalerosional processes will lead to the production of large amounts of plastic micro and nano-particles(Cozar et al., 2014; Andrady, 2015). Surveys have shown increasing amounts of plastic ‘marine litter’on the seafloor from coastal areas to deep basins and the Arctic (Pham et al., 2014; Tekman et al.,2017). On beaches, novel aggregates “plastiglomerates” have been found where plastic-containingdebris comes into contact with high temperatures (Corcoran et al., 2014).The degradation of plastics is mostly by solar ultra- violet radiation and in the oceans occursmostly in the photic zone (Andrady, 2015) and is notably temperature dependent (Andrady et al.,1998) (other mechanisms such as thermo-oxidation or hydrolysis do not readily occur in the ocean).The densification of small plastic particles by fouling organisms, ingestion and incorporation into he Silurian Hypothesis
Many radioactive isotopes that are related to anthropogenic fission or nuclear arms, have half-livesthat are long, but not long enough to be relevant here. However, there are two isotopes thatare potentially long-lived enough. Specifically, Plutonium-244 (half-life 80.8 million years) andCurium-247 (half-life 15 million years) would be detectable for a large fraction of the relevant timeperiod if they were deposited in sufficient quantities, say, as a result of a nuclear weapon exchange.There are no known natural sources of
Pu outside of supernovae.Attempts have been made to detect primordial
Pu on Earth with mixed success (Hoffmanet al., 1971; Lachner et al., 2012), indicating the rate of actinide meteorite accretion is smallenough (Wallner et al., 2015) for this to be a valid marker in the event of a sufficiently large nuclearexchange. Similarly,
Cm is present in nuclear fuel waste and as a consequence of a nuclearexplosion.Anomalous isotopic ratios in elements with long-lived radioactive isotopes are also possiblesignatures, for instance, lower than usual
U ratios, and the presence of expected daughterproducts, in uranium ores in the Franceville Basin in the Gabon have been traced to naturallyoccurring nuclear fission in oxygenated, hydrated rocks around 2 Ga (Gauthier-Lafaye et al., 1996).
The Anthropocene layer in ocean sediment will be abrupt and multi-variate, consisting of seeminglyconcurrent specific peaks in multiple geochemical proxies, biomarkers, elemental composition, andmineralogy. It will likely demarcate a clear transition of faunal taxa prior to the event compared he Silurian Hypothesis
The summary for the Anthropocene fingerprint above suggests that similarities might be foundin (geologically) abrupt events with a multi-variate signature. In this section we review a partialselection of known events in the paleo-record that have some similarities to the hypothesized eventualanthropogenic signature. The clearest class of event with such similarities are the hyperthermals,most notably the Paleocene-Eocene Thermal Maximum (56 Ma) (McInerney & Wing, 2011), butthis also includes smaller hyperthermal events, ocean anoxic events in the Cretaceous and Jurassic,and significant (if less well characterized) events of the Paleozoic. We don’t consider of events(such as the K-T extinction event, or the Eocene-Oligocene boundary) where there are very clearand distinct causes (asteroid impact combined with massive volcanism(Vellekoop et al., 2014),and the onset of Antarctic glaciation(Zachos et al., 2001) (likely linked to the opening of DrakePassage(Cristini et al., 2012)), respectively). There may be more such events in the record butthat are not included here simply because they may not have been examined in detail, particularlyin the pre-Cenozoic.
The existence of an abrupt spike in carbon and oxygen isotopes near the Paleocene/Eocenetransition (56 Ma) was first noted by Kennett & Stott (1991) and demonstrated to be global byKoch et al. (1992). Since then, more detailed and high resolution analyses on land and in theocean have revealed a fascinating sequence of events lasting 100–200 kyr and involving a rapidinput (in perhaps less than 5 kyr (Kirtland Turner et al., 2017)) of exogenous carbon into the he Silurian Hypothesis ◦ C(derived from multiple proxies (Tripati & Elderfield, 2004)), and there was a negative spike incarbon isotopes ( > (cid:104) ), and decreased ocean carbonate preservation in the upper ocean. Therewas an increase in kaolinite (clay) in many sediments (Schmitz et al., 2001), indicating greatererosion, though evidence for a global increase is mixed. During the PETM 30–50% of benthicforaminiferal taxa became extinct, and it marked the time of an important mammalian (Aubryet al., 1998) and lizard (Smith, 2009) expansion across North America. Additionally, many metalabundances (including V, Zn, Mo, Cr) spiked during the event (Soliman et al., 2011). In the 6 million years following the PETM, there are a number of smaller, though qualitativelysimilar, hyperthermal events seen in the record (Slotnick et al., 2012). Notably, the Eocene ThermalMaximum 2 event (ETM-2), and at least four other peaks are characterized by significant negativecarbon isotope excursions, warming and relatively high sedimentation rates driven by increasesin terrigenous input (D’Onofrio et al., 2016). Arctic conditions during ETM-2 show evidence ofwarming, lower salinity, and greater anoxia (Sluijs et al., 2009). Collectively these events have beendenoted Eocene Layers of Mysterious Origin (ELMOs) .Around 40 Ma, another abrupt warming event occurs (the Mid-Eocene Climate Optimum(MECO)), again with an accompanying carbon isotope anomaly (Galazzo et al., 2014). First identified by Schlanger & Jenkyns (1976), ocean anoxic events (OAEs), identified by periodsof greatly increased organic carbon deposition and laminated black shale deposits, are times whensignificant portions of the ocean (either regionally or globally) became depleted in dissolved oxygen,greatly reducing aerobic bacterial activity. There is partial (though not ubiquitous) evidence duringthe larger OAEs for euxinia (when the ocean column becomes filled with hydrogen sulfide (H S)) While it is tempting to read something into the nomenclature of these events, it should be remembered thatmost things that happened 50 million years ago will forever remain somewhat mysterious. he Silurian Hypothesis δ C(as in the PETM), followed by a positive recovery during the events themselves as the burial of(light) organic carbon increased and compensated for the initial release (Jenkyns, 2010; Naafset al., 2016; Mutterlose et al., 2014; Kuhnt et al., 2011). Causes have been linked to the crustalformation/tectonic activity and enhanced CO (or possibly CH ) release, causing global warmth(Jenkyns, 2010). Increased seawater values of Sr/ Sr and
Os/
Os suggest increased runoff,greater nutrient supply and consequently higher upper ocean productivity (Jones, 2001). Possiblehiatuses in some OAE 1a sections are suggestive of an upper ocean dissolution event (Bottini et al.,2015).Other important shifts in geochemical tracers during the OAEs include much lower nitrogenisotope ratios ( δ N), increases in metal concentrations (including As, Bi, Cd, Co, Cr, Ni, V)(Jenkyns, 2010). Positive shifts in sulfur isotopes are seen in most OAEs, with a curious exceptionin OAE-1a where the shift is negative (Turchyn et al., 2009). he Silurian Hypothesis Starting from the Devonian period, there have been several major abrupt events registered interrestrial sections. The sequences of changes and the comprehensiveness of geochemical analysesare less well known than for later events, partly due to the lack of existing ocean sediment, butthese have been identified in multiple locations and are presumed to be global in extent.The Late Devonian extinction around 380–360 Ma, was one of the big five mass extinctions.It’s associated with black shales and ocean anoxia (Algeo & Scheckler, 1998), stretching from theKellwasser events ( ∼
378 Ma) to the Hangenberg event at the Devonian-Carboniferous boundary(359 Ma) (Brezinski et al., 2009; Vleeschouwer et al., 2013).In the late Carboniferous, around 305 Ma the Pangaean tropical rainforests collapsed (Sahneyet al., 2010). This was associated with a shift toward drier and cooler climate, and possibly areduction in atmospheric oxygen, leading to extinctions of some mega-fauna.Lastly, the end-Permian extinction event (252 Ma) lasted about 60 kyr was accompanied by aninitial decrease in carbon isotopes (-5–7 (cid:104) ), significant global warming and extensive deforestationand wildfires (Krull & Retallack, 2000; Shen et al., 2011; Burgess et al., 2014) associated withwidespread ocean anoxia and euxinia (Wignall & Twitchett, 1996). Pre-event spikes in nickel (Ni)have also been reported (Rothman et al., 2014).
There are undoubted similarities between previous abrupt events in the geological record and thelikely Anthropocene signature in the geological record to come. Negative, abrupt δ C excursions,warmings, and disruptions of the nitrogen cycle are ubiquitous. More complex changes in biota,sedimentation and mineralogy are also common. Specifically, compared to the hypothesizedAnthropocene signature, almost all changes found so far for the PETM are of the same sign andcomparable magnitude. Some similarities would be expected if the main effect during any eventwas a significant global warming, however caused. Furthermore, there is evidence at many of theseevents that warming was driven by a massive input of exogeneous (biogenic) carbon, either as CO or CH . At least since the Carboniferous (300–350 Ma), there has been sufficient fossil carbon to he Silurian Hypothesis or CH to the atmosphere. Impacts to warming and/orcarbon influx (such as increased runoff, erosion etc.) appear to be qualitatively similar whenever inthe geological period they occur. These changes are thus not sufficient evidence for prior industrialcivilizations.Current changes appear to be significantly faster than the paleoclimatic events (figure 1), butthis may be partly due to limitations of chronology in the geological record. Attempts to time thelength of prior events have used constant sedimentation estimates, or constant-flux markers (e.g. He (McGee & Mukhopadhyay, 2012)), or orbital chronologies, or supposed annual or seasonalbanding in the sediment (Wright & Schaller, 2013). The accuracy of these methods suffer whenthere are large changes in sedimentation or hiatuses across these events (which is common), orrely on the imperfect identification of regularities with specific astronomical features (Pearson &Nicholas, 2014; Pearson & Thomas, 2015). Additionally, bioturbation will often smooth an abruptevent even in a perfectly preserved sedimentary setting. Thus the ability to detect an event onsetof a few centuries (or less) in the record is questionable, and so direct isolation of an industrialcause based only on apparent timing is also not conclusive.The specific markers of human industrial activity discussed above (plastics, synthetic pollutants,increased metal concentrations etc.) are however a consequence of the specific path human societyand technology has taken, and the generality of that pathway for other industrial species is totallyunknown. Large-scale energy harnessing is potentially a more universal indicator, and given thelarge energy density in carbon-based fossil fuel, one might postulate that a light δ C signal mightbe a common signal. Conceivably, solar, hydro or geothermal energy sources could have beentapped preferentially, and that would greatly reduce any geological footprint (as it would ours).However any large release of biogenic carbon whether from methane hydrate pools or volcanic he Silurian Hypothesis he Silurian Hypothesis
No funding has been provided nor sought for this study. We thank Susan Kidwell for being generouswith her time and helpful discussions, David Naafs and Stuart Robinson for help and pointers to datafor OAE1a and Chris Reinhard for his thoughtful review. The GISTEMP data in Fig. 1a was from https://data.giss.nasa.gov/gistemp (accessed Jul 15 2017).
REFERENCES
Achille, G. D. & Hynek, B. M. 2010, Nature Geoscience, 3, 459Algeo, T. J. & Scheckler, S. E. 1998, Philosophical Transactions of the Royal Society B: BiologicalSciences, 353, 113Andrady, A., Hamid, S., Hu, X., & Torikai, A. 1998, Journal of Photochemistry and PhotobiologyB: Biology, 46, 96 he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis he Silurian Hypothesis
Figure 1.
Illustrative stable carbon isotopes and temperature (or proxy) profiles across three periods. a) Themodern era (from 1600 CE with projections to 2100). Carbon isotopes are from sea sponges (Böhm et al., 2002), andprojections from Köhler (2016). Temperatures are from Mann et al. (2008) (reconstructions), GISTEMP (Hansenet al., 2010) (instrumental) and projected to 2100 using results from Nazarenko et al. (2015). Projections assumetrajectories of emissions associated with RCP8.5 (van Vuuren et al., 2011). b) The Paleocene-Eocene ThermalMaximum (55.5 Ma). Data from two DSDP cores (589 and 1209B) (Tripati & Elderfield, 2004) are used to estimateanomalous isotopic changes and a loess smooth with a span of 200 kya is applied to make the trends clearer.Temperatures changes are estimated from observed δ O carbonate using a standard calibration (Kim & O’Neil, 1997).c) Oceanic Anoxic Event 1a (about 120 Ma). Carbon isotopes are from the La Bédoule and Cau cores from thepaleo-Tethys (Kuhnt et al., 2011; Naafs et al., 2016) aligned as in Naafs et al. (2016) and placed on an approximateage model. Data from Alstätte (Bottini & Mutterlose, 2012) and DSDP Site 398 (Li et al., 2008) are aligned basedon coherence of the δ13