Three flaws in defining a formal ‘Anthropocene’

I Introduction

Few (if any) scientists question the concept that an anthropogenic era now exists – an interval during which human activities have exceeded natural forces in altering many aspects of Earth’s land surfaces, oceans, and atmospheric composition. Nature still plays an important role, especially during disruptive storms, volcanic eruptions, earthquakes, and other disasters, as well as in gradual processes such as tectonic activity, tidal actions, and orbital changes, but for many processes (especially biogeochemical ones), human activities dominate (Steffen et al., 2004). Consequently, the proposal of Crutzen and Stoermer (2000) that we live in an ‘anthropogenic age’ altogether different from the world preceding it has received well-deserved attention in the sciences, the humanities, and many public venues.

An effort is now underway to give this informal acknowledgement of human influence formal scientific recognition. In 2016, the Anthropogenic Working Group (AWG) of the subcommission of Quaternary Stratigraphy of the Geological Society of London put forward a preliminary recommendation to define a formal ‘Anthropocene’ epoch that began on July 16, 1945 with the first atomic bomb detonation at Alamagordo, New Mexico. Although this recommendation was returned by the International Commission on Stratigraphy (ICS) for further consideration, the prevailing AGW view still favors a mid/late-1900s choice (Waters et al., 2017; Zalasiewicz et al., 2017b).

For a formal AWG recommendation to become official, it must first be endorsed by the Subcommission of Quaternary Stratigraphy of the ICS, then accepted by the full ICS voting membership, and, finally, ratified by the Executive Committee of the International Geological Congress. To be acceptable, several criteria (formal rules) must be met. Evidence must be shown of major environmental changes that are globally synchronous and traceable in Earth’s biogeochemical records on large spatial scales. In addition, the level in geologic records chosen to mark the start of the proposed interval must be based on observable rock properties at a type locality, with a lower boundary marked by a ‘golden spike’. The methodology is available at http://www.stratigraphy.org/index.php/ics-chart-timescale and has been briefly summarized by Finney and Edwards (2016). The acceptance of a new designation in the International Chronostratigraphic Chart ultimately becomes the basis for inclusion in the Geologic Time Scale.

At first glance, choosing an Anthropocene start in the middle to late 1900s seems to meet these criteria. By then, cumulative alterations of the environment by humans had become enormous, and by many measures had begun to accelerate rapidly (Steffen et al., 2015; Zalasiewicz et al., 2017b). In addition, the selection of the start of nuclear testing would provide an isotopic marker detectable in many proxy recorders of Earth’s climatic history. This 1900s start date for a formal Anthropocene would then supersede the initial suggestion by Crutzen and Stoermer (2000), who had proposed that it began in the early–middle 1800s during the early part of the industrial era.

Other scientists have proposed start dates well before the industrial era (Certini and Scalenghe, 2011; Glikson, 2013; Kunnas, 2017; Lewis and Maslin, 2015; Smith and Zeder, 2013). Still, other scientists have argued against any formal definition of the anthropogenic era (Edgeworth et al., 2015; Ruddiman, 2013a; Ruddiman et al., 2015; Rull, 2017). This paper focuses on the arguments of this latter group.

Section II of this paper summarizes the case that a formally designated late-1900s ‘Anthropocene’ start date would ignore wide-ranging evidence for the vast transformations of the surface of this planet long before the industrial era (Ellis et al., 2013; Ruddiman, 2013a; Ruddiman et al., 2015). These alterations, including major extinctions of marsupials and mammals, widespread deforestation of most regions where rain-fed agriculture is practiced today, and the more recent conversion of semi-arid lands to agriculture, are far too large to be relegated to the ‘pre-Anthropocene’.

Section III makes the point that the ‘rules’ under which the AWG operates do not permit the recognition of these early changes as part of the Anthropocene, no matter how extensive they are. The marsupial and later mammal extinction pulses fail to qualify because they were constrained to one or two continents and thus lacked sufficient spatial extent. The spread of agriculture across all of Earth’s arable regions is rejected because it was time-transgressive and cannot be assigned a single start date. These formal rules do not just constrain the method chosen by the AWG, they render it inadequate to the task.

Section IV provides an assessment of the relevance of the formal method relied on by the AWG to scientists actively examining the younger geologic record. For centuries, the development of geoscience relied on this method to subdivide the Geologic Column, an historic achievement. In recent decades, however, most formal designations of this kind for deposits in the younger part of the geologic record (the Pleistocene Epoch) have been widely ignored by practicing scientists. Instead, geochemical dating, combined with associated methods such as magnetic stratigraphy and orbital tuning, have been the standard method of dating sequences.

II Flaw 1: a 1900s date would ignore extensive earlier changes

Pre-industrial alterations of planet Earth by humans cover a wide range of changes (Figure 1). The extinction of a majority of large marsupials in Australia near 50,000 years ago was, within dating uncertainties, coincident with the first arrival of humans. The absence of major changes in regional Australian climate at that time implicates humans as responsible (Barnosky et al., 2004). A second extinction pulse, in the Americas near 12,000 years ago, killed off 80% of the large-mammal genera, decimating what had been a diverse Serengeti-like environment in the western American plains, with similar losses in South America. Although this extinction pulse occurred during a rapid deglacial climate change, numerous previous deglaciations with similarly rapid warming and ice-sheet melting had produced no such extinction event, again strongly implicating the role of indigenous people (Barnosky et al., 2004).

OPEN IN VIEWER

Figure 1.

The industrial era has been a time of greatly accelerated environmental changes caused by humans, but it was preceded by major earlier anthropogenic changes (from Ruddiman et al., 2015). These earlier changes included massive pulses of mammal extinctions in Australia and the Americas; changes associated with agriculture, including the spread of domesticated crops and livestock, forest clearance, soil erosion, irrigation of rice paddies, and substantial anthropogenic emissions of CO₂ and CH₄ to the atmosphere. None of these far-ranging early anthropogenic changes would be included in a formally defined ‘Anthropocene’ starting in the middle–late 1900s, as now being considered by the AWG.

By 12,000 to 8000 years ago, humans had begun to domesticate crops and livestock in several regions, and these innovations slowly spread across Earth’s arable regions over many millennia (Figure 2). Agriculture was introduced both by movements of people carrying new technologies and by converting hunter-gatherers to new skills.

OPEN IN VIEWER

Figure 2.

Domesticated crops and livestock originated in several founder areas from 12,000 to 4200 years ago. Subsequent dispersion pathways are shown by arrows with ages in thousands of years (from Fuller et al., 2014; Larson et al., 2014).

Until recently, the major early anthropogenic change resulting from the spread of agriculture – clearance of natural forests – was considered as primarily an industrial-era phenomenon (Joos et al., 2004; Pongratz et al., 2009; Stocker et al., 2011; Strassmann et al., 2008), but evidence from three independent sources refutes this conclusion.

Ground-truth evidence of early forest clearance in north-central Europe comes from compilations of pollen trends by Fyfe et al. (2014), Gaillard et al. (2010), and Roberts et al. (2018). They found that Europe was largely wooded 6000 years ago, but a long-term trend toward reduced forest cover that began 5000 years ago had reached modern levels (scattered forest patches and woodlands) centuries before the industrial era. In north-central China, extensive archeological investigations reported by Li et al. (2009) show a very rapid 30-fold expansion of archaeological settlements between 8000 and 4000 years ago because of farming. The most detailed provincial records place the first expansion of farming settlements at 7000 years ago (Hosner et al., 2016). These two regions – Europe and China – are the only ones studied in sufficient detail to provide broad overviews of early farming and forest clearance, but many regional studies point to similar trends well before the industrial era (Ellis et al., 2013; Ruddiman, 2013b).

A second line of evidence for extensive early clearance comes from the land-use simulation of Kaplan et al. (2011), based on historical data that show large per-capita clearance millennia ago compared to much smaller levels by the 1600s and 1700s. They estimated global carbon emissions from vegetation clearance during pre-industrial times at 343 GtC (1 Gt = 1 billion tons), well above the industrial-era total.

The Kaplan et al. (2011) simulation is supported by a third kind of estimate – the assessment of Earth’s potential vegetation carbon stocks in the absence of human influence (Erb et al., 2017). Erb and colleagues estimated a potential modern vegetation total of 916 GtC, compared to the modern biomass of 450 GtC, leaving the difference of 466 GtC as the amount removed by anthropogenic land use. With 156 GtC of this land-use removal having occurred during the industrial era (Houghton, 2007), the remaining residual of 310 GtC is the estimated amount of pre-industrial deforestation and other vegetation changes.

These estimates of 343 and 310 GtC closely match the rough (back-of-the envelope) 300 GtC proposed by Ruddiman (2003). These values indicate that pre-industrial deforestation was twice or more the 156 GtC emitted in the post-1850 industrial era (Houghton, 2007). And if the AWG chooses 1950 for the start of the Anthropocene, the additional ∼75 Gt of carbon emissions that occurred between 1850 and 1950 would bring total carbon emissions from deforestation before 1950 to ∼400 Gt, compared to just ∼75 Gt since 1950.

This extent of pre-industrial forest clearance would have significantly altered the composition of Earth’s atmosphere (using the standard 14.2 GtC = 1 ppm CO₂ formulation). Although the earlier simulations of small early clearance noted above would only have increased the atmospheric CO₂ concentration by just a few ppm before the industrial era, the recent evidence is more in line with the proposal by Ruddiman (2003) of an anthropogenic CO₂ increase measured in tens of ppm since 7000 years ago. Emissions of 300–340 GtC would have caused CO₂ to increase by 21–24 ppm. The lack of any such increase in previous interglaciations (Ruddiman et al., 2013a) supports the case for substantial anthropogenic emissions since 7000 years ago, along with consequent effects in ameliorating a natural cooling trend toward glacial conditions.

A second major pre-industrial change of Earth’s surface was the spread of irrigated rice farming across southern and southeastern Asia beginning 5000 years ago (Fuller et al., 2011). The irrigated rice paddies replaced what had once been natural wetlands that were drying out under a weakening summer monsoon, and they also extended these anthropogenic wetlands into regions that had previously been dry land, such as terraced hillsides. After 5000 years ago, domesticated livestock that emitted methane also spread across rain-fed regions of Asia, Europe, and Africa (Fuller et al., 2011).

A third extensive alteration of Earth’s surface began early in the industrial era with the invention of the John Deere plow, the McCormick reaper, and, soon afterwards, tractors powered by fossil fuels. These inventions, along with irrigation water drawn from deep aquifers, converted much of Earth’s semi-arid steppes and prairies once covered by grasses, herbs, and forb vegetation into geometrically patterned croplands.

These changes altered Earth’s surface extensively. Together, they are by far the largest transformations of Earth’s surface in all of human history. In comparison, the rapid expansion of urban areas (now ∼1% of the total), the developing networks of roads, the present-day regional deforestation in Brazil and Indonesia, and other recent changes are second-order.

Given that the pre-1950 changes itemized above have been the largest transformations of Earth’s surface in all of human history, how can they possibly be excluded from a formal ‘Anthropocene’ by assigning a start date in the middle or late 1900s? As noted by Ruddiman et al. (2015),

Does it really make sense to define the start of a human-dominated era millennia after most forests in arable regions had been cut for agriculture, most rice paddies had been irrigated, and CO₂ and CH₄ concentrations had been rising because of agricultural and industrial emissions?

In summary, the AWG strategy would relegate the vast pre-industrial (and pre-1950s) alterations of this planet to the ‘pre-Anthropocene’. These alterations were the largest transformations of Earth’s surface to date by human activities. How does this make sense?

III Flaw 2: the AWG method is unable to acknowledge early changes

What can explain this confusing, and apparently looming, decision by the AGW to shunt all these extensive early changes into the ‘pre-Anthropocene’? The basic reason is embedded in the inherent constraints of the method they have chosen, which developed historically from the discipline of geology. This field grew out of contributions during past centuries from individuals known as ‘natural philosophers’ and, later, ‘scientists’. Prominent examples include: Nicolas Steno, James Hutton, William ‘Strata’ Smith, and Charles Lyell. By meticulous fieldwork, observations, and deduction, these pioneers developed the principles which formed the basis of geological science, notably: superposition of strata, reliance on changing fossil sequences to establish relative ages of strata, recognition of complications from facies changes, and use of index fossils to overcome those complications. This centuries-long effort led to the construction of the standard Geologic Column, which integrates all investigations into a single representative record extending from the oldest known strata to the youngest, all ordered according to the principles just noted.

The formal structure created by these efforts is, however, inadequate to the task of defining a formal ‘Anthropocene’ because the principles (‘rules’) for adding new divisions to the Geologic Column do not permit the inclusion of any of the extensive pre-industrial changes described in Section II. Because the highly concentrated pulse of large-marsupial extinctions was localized to Australia and the pulse of rapid large-mammal extinctions to the Americas, both changes, even though profound and ecologically transformative for those continents, lack the requisite global or near-global spatial extent to qualify for inclusion in a formal ‘Anthropocene’.

The spread of agriculture does qualify as a near-global phenomenon but the AWG rejects it as part of a formal ‘Anthropocene’ because it is time-transgressive, with different times of initial domestication of crops and livestock in the various founder areas and different times of arrival as these discoveries spread to other regions (Figure 2). Zalasiewicz et al. (2017a) state, ‘It is this gradual nature of land-use change…that makes it such a poor indicator for the commencement of an isochronous Anthropocene Epoch’. Waters et al. (2017) state, ‘In an Anthropocene context, the choice of markers should not be to provide an indication of the start of anthropogenically driven effects, but to provide the most pragmatic marker that will allow global correlation of the chosen boundary’. These statements clearly reveal the a priori bias that lies at the heart of the AWG approach. Rather than recognize these vast early transformations of Earth’s surface as a fully legitimate part of the anthropogenic era, the AWG favors holding to formal rules that shunt these extensive changes to ‘pre-Anthropocene’ times.

Some scientists (Edgeworth et al., 2015; Ruddiman, 2013a; Ruddiman et al., 2015) agree that the onset and spread of agriculture were time-transgressive and cannot be assigned a single start date. Ruddiman (2013a) and Ruddiman et al. (2015) have been interpreted as proposing the start of a formal ‘Anthropocene’ in the early–middle Holocene with the rise of agriculture (Rull, 2017; Zalasiewicz et al., 2017b), but this interpretation is incorrect. To the contrary, those papers argued against any formal definition of the Anthropocene precisely because of its basic time-transgressiveness.

Ruddiman et al. (2015) suggested that the ‘anthropocene’ be treated informally (with a small ‘a’), thus giving scientists the flexibility to add specific modifiers as needed: the ‘early agricultural anthropocene’, the ‘pre-industrial anthropocene’, the ‘industrial-era anthropocene’, the ‘post-nuclear (late 1900s) anthropocene’, and others. This informal approach would allow the use of the term to capture the full rich history of human influences on this planet, rather than just the accelerating changes since 1950.

This approach would, however, require the AWG to admit that their current formal methodology is inappropriate for the task at hand and to abandon it. In view of the fact that their chosen method for defining the start of the formal (capital-A) ‘Anthropocene’ is unable to acknowledge explicitly either the extinctions of marsupials and other mammals or the spread of agriculture, how could they not come to this conclusion?

The AWG preference for a formal approach may, in part, reflect a natural human tendency to want to impose formal definitions and names on complex processes. This formalisitic approach has been beneficial to many scientific endeavors, but its utility in this case is flawed.

Another explanation for this preference may be that the AWG method emerged from the discipline of geology, and most of the 37 members of the working group are geologists or geochemists with little or no direct experience in studying human activities that occurred long before the industrial era (Ellis et al., 2013). The two AWG archaeologists (Matt Edgeworth and Cath Neal) are on record as favoring a diachronous small-‘a’ anthropocene (Edgeworth et al., 2015). Would more scientists from other disciplines have reached similarly different conclusions?

In summary, a second flaw in the AWG approach is its inability to admit that the large but time-transgressive agricultural changes in pre-industrial time are worthy of inclusion in a formal ‘Anthropocene’. Faced with a choice between fully acknowledging the vast spatial extent of these early changes, or continuing to adhere to rules developed for other purposes centuries ago, the AWG has (to date) held on to their ‘rules’.

IV Flaw 3: many scientists will likely ignore any formal designation

If the AWG goes ahead and recommends the formal recognition of an ‘Anthropocene’ beginning in the middle–late 1900s, and if it is approved in following actions, what use are practicing scientists likely to make of this initiative? Clues to their reaction can be found in the extent to which they have recently made use of the subdivisions of the Pleistocene Epoch that have been defined by formal stratigraphic rules.

Throughout the centuries when the construction of the Geologic Column was underway, the issue of how much time it encompassed was an ongoing uncertainty. Scientists like Hutton and Lyell working face-to-face with the rock record in the field saw evidence that the age of the Earth was immense. They realized that processes now at work on Earth’s surface could account for features in the rock record if acting over enormous amounts of time. Still, estimates of elapsed time varied widely.

The 1950s saw increasingly accurate assessments of Earth’s temporal history by dating strata and/or their interbedded igneous rocks by the gradual decay of radioactive isotopes compared to their stable, non-radioactive, non-decaying cousins. For older strata, measurements of U/Pb, K/Ar, and Ar/Ar now provide dates that are often accurate to within 1%. For deposits thousands to tens of thousands of years in age, ¹⁴C dates (adjusted to Th/U corrections) can provide similar accuracy.

For most of the older geologic record, geoscientists still refer to the formally defined Epochs and Stages to provide an initial sense of the time interval under study, but they usually cite radiometric dating for accurate time control. For the younger part of the geologic record, however, the emphasis is different. For example, the term ‘Pleistocene’ is widely used because it refers to the time when large ice sheets advanced and retreated across Northern Hemisphere continents. In 1948, the 18th International Geological Congress recommended that the start of the Pleistocene Epoch be placed at 1.66 million years, based on the first appearance of the cool-water mollusks Arctica islandica and Hyalinea baltica in a sedimentary section in Sicily. More than three decades later, the 27th International Geological Congress in 1985 formally adopted a slightly altered version of this proposal with an age of 1.8 million years.

During those decades, however, evidence accumulated that large northern hemisphere ice sheets were present well before either 1.66 or 1.8 million years ago. Decisively, North Atlantic sediment sequences dated by magnetic stratigraphy and associated Ar/Ar dating (Shackleton et al., 1984) showed that major glaciations and deposition of ice-rafted debris began around 2.7 million years ago, near the Gauss/Matuyama paleomagnetic boundary.

In 2009, in a long-delayed response to this and other evidence of earlier cooling, the ICS updated the start of the Pleistocene to 2.588 million years ago. Partridge (1997) and Gibbard and Head (2009) describe the long process of arriving at this formal decision. Meanwhile, most paleoclimatologists had simply ignored this ongoing debate and adopted an age of 2.7 or 2.6 million years for the start of the Pleistocene ice-age cycles based on the North Atlantic evidence from Shackleton et al. (1984). Practicing scientists were not paying much attention to the slow process of formally approved designations.

More strikingly, the four formally defined stratigraphic stages of the Pleistocene are almost never mentioned in the research literature. Instead, radiometric dating combined with magnetic stratigraphy and cyclostratigraphy (orbital tuning) now prevail.

The stages that now formally subdivide the Pleistocene Epoch are as follows: Gelasian (2.588 to 1.8 million years), Calabrian (1.8 to 0.781 million years), Ionian (0.781 to 0.126 million years), and Tarantian (0.126 to 0.00117 million years). But these formally designated names are rarely used. Most books and textbooks on paleoclimate and paleoenvironment in recent decades (some with several editions) do not even mention these formally approved stage divisions: Imbrie and Imbrie (1979); Hecht (1983); Crowley and North (1991); Bradley (1999); Kump et al. (1999); and Ruddiman (2001). Cronin (2010) includes the stages in a table early in his book ‘Paleoclimates’, but does not subsequently mention them anywhere in the text. Despite my 50 years of experience in paleoclimatology, I had to look up these stage names to be able to refer to them correctly in this paper.

Many peer-reviewed publications covering this time interval make use of informal (although sometimes capitalized) terms such as the ‘mid-Pleistocene transition’ from dominant 41,000-year cycles to 100,000-year oxygen-isotopic cycles, but they invariably ignore the formally designated geologic stages. The primary reason for this is obvious in Figure 3: the four formally designated stages do not mark any obvious transitions in the climatically diagnostic benthic foraminiferal δ¹⁸O signal, which shows a gradual trend toward more positive values (greater ice volume and colder deep-ocean temperatures) with orbital-scale cycles superimposed. For most scientists, the formally designated Pleistocene stages are thus of no practical use in this well-dated continuum of gradual cooling.

OPEN IN VIEWER

Figure 3.

Comparison of the Pleistocene stages formally defined in 2009 by the International Commission on Stratigraphy with the stacked benthic δ¹⁸O record from Lisiecki and Raymo (2005). From Figure 11 in Ellis (2018), used with permission.

This disregard by practicing scientists for the formal Pleistocene stage designations may apply to any future attempt to name a formal ‘Anthropocene’ with a start date in the middle–late 1900s. Scientists who explore post-1950 phenomena may well find it convenient, but those who investigate anthropogenic changes during the millennia, centuries, and even decades before the middle–late 1900s are unlikely to be enthusiastic about a definition that relegates their studies to ‘pre-Anthropogenic’ time.

We live in an era of radiometric dating. Increasingly sophisticated geochemical methods have gradually made the classic process of assigning formal geologic subdivisions obsolete for younger strata, especially those of the Pleistocene Epoch. In addition to radiometric dating, many records lie within the reach of historical records (the last two or three millennia) and can be assigned to specific Gregorian calendar years. I suspect that most practicing scientists will refer to the ‘anthropocene’ and its various phases in an informal way and ignore any formal late-1900s ‘Anthropocene’ designation because of its inflexibility or inapplicability to the records they study.

In summary, a third flaw in the AWG strategy is its reliance on an antiquated method that has for decades been widely ignored by scientists working on Pleistocene-age records and does not apply to the very wide range of pre-1900s studies of anthropogenic phenomena.

V Summary

Any proposal to formally define an Anthropocene beginning in the middle or late 1900s would have several flaws. First, it would omit much of the rich history of how our species has transformed this planet, by killing off large marsupials in Australia and large mammals in the Americas and by introducing agriculture and extending it to Earth’s arable areas. Second, the methods (‘rules’) used by the AWG are flawed because they are fundamentally unable to recognize these earlier transformations, no matter how massive. Third, the process used by the working group (formal definition of subdivisions of the geologic column) is now largely obsolete for younger geologic records, having been replaced by radiometric dating and associated techniques. Finally, this potential AWG recommendation would constrain the use of the term ‘Anthropocene’ to a single rigid definition, rather than using the informal term ‘anthropocene’, which is open to a wide range of flexible applications. My arguments do not in any way reject the large acceleration of many anthropogenic influences in the last several decades, but these recent changes are no reason for excluding all the profound early changes by choosing a formal definition.