Comment: How long have humans been altering Earth's climate?

by William F. Ruddiman
Thursday, April 7, 2016

The early anthropogenic hypothesis holds that prolonged warmth in the Holocene was caused by early agriculture. Paleoclimatologists have long sought an analog of the current Holocene interglaciation. Here, precession (A) and obliquity (B), which impact solar insolation, and atmospheric carbon dioxide concentrations (C) of previous interglaciations are compared with the Holocene (stage 1). The closest analog to the Holocene is stage 19, which began about 787,000 years ago. Credit: Ruddiman et al., Review of Geophysics, 2016.

Throughout the 20th century the paleoclimate science community regarded the warmth of the current (Holocene) interglaciation — prior to the major anthropogenic intervention of the last 150 years — as overwhelmingly natural in origin. In this view, changes in Earth’s orbit had resulted in increased summer insolation at high latitudes of the Northern Hemisphere, which began melting the North American and Scandinavian ice sheets 17,000 years ago and eventually ushered in interglacial warmth. Although further orbital shifts then led summer insolation to begin decreasing 10,000 years ago, the drop was thought not to have been substantial enough to cause renewed glaciation.

In the “early anthropogenic hypothesis,” first published in 2003, I proposed a different interpretation: that greenhouse gas emissions from early agriculture were the main reason for prolonged warmth lasting into modern times. I noted that in three previous interglaciations observed in ice-core records, concentrations of both carbon dioxide and methane decreased through the first 10,000 years, resulting in cooling trends that led to renewed glaciation. But the equivalent part of the current interglaciation has been different: Carbon dioxide concentrations rose during the last 7,000 years, and methane concentrations rose during the last 5,000 years. I proposed that these anomalous greenhouse gas increases were anthropogenic in origin and kept global climate warmer than it would have been in a world controlled only by nature.

I attributed the anomalous rise in carbon dioxide since 7,000 years ago to early deforestation, and the rising methane trend since 5,000 years ago to the spread of rice irrigation and livestock tending. Back in 2003, quantitative information about early agriculture was scarce, except for one very interesting data point: The 1086 Domesday Book, a survey ordered by William the Conqueror, reported that forests covered just 15 percent of Britain, indicative of early deforestation, with forest cover already similar, in fact, to levels today.

The early anthropogenic hypothesis was received with acclaim by some scientists, but with deep skepticism by others. From 2004 to 2009, several prominent climate scientists published papers criticizing it. The most prevalent argument was that far too few people were living thousands of years ago to have caused land-use changes sufficient to alter greenhouse gas levels.

Another criticism centered on a previous interglaciation that was proposed as the closest orbital analog to the current interglacial — stage 11, which began about 424,000 years ago. That interglaciation was thought to have lasted 26,000 years, compared to just 11,000 years (to date) for the present one, suggesting that some 15,000 years of interglacial warmth potentially still remain ahead of us now before the next glaciation.

In addition, ratios of carbon-13 to carbon-12 in ice-core carbon dioxide, an index of net global terrestrial emissions, showed only a weak decrease since early in the current interglaciation. This implied very low deforestation emissions during the last 7,000 years, equivalent to an addition of just 2 to 3 parts per million (ppm) carbon dioxide to the atmosphere.

Since 2010, however, a surge of new evidence converging from a wide range of geoscience-related disciplines has refuted these criticisms and lent support to the early anthropogenic hypothesis. Surveys of historical records by ecological modeler Jed Kaplan, myself and landscape ecologist Erle Ellis found that farmers in both Europe and China 2,000 years ago used at least four times as much land per person as those in the centuries just before the industrial era (the 1700s). Early slash-and-burn farming practices that rotated from plot to plot were highly inefficient and cleared large amounts of land. In contrast, modern farmers plant one or more crops on the same land every year.

This evidence of early farmers clearing more land per capita suggests, for example, that the 200 million to 250 million people living 2,000 years ago were using an area of land for agriculture equivalent to that used by almost a billion people in more recent times. Based on this evidence, Kaplan ran a land-use simulation that found that preindustrial anthropogenic carbon dioxide emissions would have added 24 ppm to the atmospheric concentration, five times the amount estimated by modelers who had assumed that early farmers used relatively small amounts of land through the millennia.

Work by geologist and palynologist Chronis Tzedakis and others showed that interglacial stage 11 is not a good orbital analog for the current interglaciation after all. While both interglaciations were characterized by similar low-amplitude changes in the eccentricity-modulated precession of Earth’s orbit, the obliquity of the orbit during stage 11 was far offset from the one in the current interglaciation (see figure parts A and B).

The optimal orbital analog to the Holocene turns out to be interglacial stage 19, which began about 787,000 years ago. Comparison shows that the orbital indices of each are almost exactly in phase (see figure parts A and B). Yet, during the first 10,000 years of stage 19, atmospheric carbon dioxide concentrations fell steadily to between 240 and 245 ppm. This is identical to the range predicted by the early anthropogenic hypothesis (see figure part C) for a Holocene interglaciation free of an early anthropogenic overprint.

Archaeobotanist Dorian Fuller and his colleagues mapped the spread of irrigated rice across southern Asia from 5,000 to 1,000 years ago. He found that emissions from methane-producing rice paddies accounted for 70 parts per billion (ppb) out of the 100-ppb methane increase observed in ice cores during that interval. He also mapped the spread of methane-emitting livestock, but has not estimated their methane emissions.

Paleoecologists Ralph Fyfe, Jessie Woodbridge and Neil Roberts and their colleagues recently published results of a comprehensive synthesis of pollen data from hundreds of cores across north-central Europe. They showed that deforestation was nearly complete before the industrial era, as I had claimed in the early anthropogenic hypothesis. In Britain, deforestation was already extensive by 2,000 years ago, consistent with the evidence from the Domesday Book. No other continent has been analyzed in comparable detail, but planning for such studies is underway.

The problem of the weak carbon-13 isotopic signal turns out to have a plausible explanation also: a coincidental offsetting of the anthropogenic emissions by natural burial of carbon in boreal peats. Ecologist Zicheng Yu has shown that the amount of carbon stored in peats during the last 7,000 years is equivalent to a 21-ppm reduction of carbon dioxide in the atmosphere. Similar increases in carbon storage in peats during previous interglaciations must have contributed to the decreases in atmospheric carbon dioxide observed in ice cores at those times, with resulting cooling effects on climate. During the current interglaciation, however, anthropogenic emissions from deforestation offset that natural trend. The 24-ppm anthropogenic carbon dioxide release due to deforestation estimated by Kaplan slightly outweighs the 21-ppm carbon dioxide withdrawal Yu estimated for carbon burial in peats and satisfies the carbon-isotopic constraint.

Another factor in the global carbon budget of the last 7,000 years is carbon dioxide feedback from the ocean. Direct emissions of carbon dioxide from deforestation represent an anthropogenic anomaly compared to previous interglaciations. These emissions would have prevented both the atmosphere and oceans from cooling, keeping them anomalously warm.

Experiments by climate modelers John Kutzbach, Steve Vavrus and Feng He suggest whole-ocean warming of 0.8 to 0.9 degrees Celsius or more, sufficient to cause two kinds of feedbacks that would have released stored carbon dioxide from the oceans. Reduced carbon dioxide solubility in the warm ocean would have emitted 6 ppm or more carbon dioxide to the atmosphere. Additionally, anomalous warmth in the Southern Ocean would have reduced sea-ice cover and prompted carbon dioxide exchange with overlying air masses, again increasing carbon dioxide in the atmosphere. These feedbacks count as anthropogenic because they are the result of direct emissions from deforestation.

One indication of the expanding impact of the early anthropogenic hypothesis on the larger scientific community is the evolving agenda of the Past Global Changes (PAGES) program, a core project of the International Geosphere-Biosphere Program. In the 1980s, PAGES began a major initiative called the “2K” project, with the goal of measuring paleoclimatic responses during the last 2,000 years in order to construct a natural baseline against which to assess the anthropogenic effects of the last 150 years.

Last year, however, PAGES began a new effort called the “Land Cover 6K” project, motivated by the growing realization that humans have been altering land cover, greenhouse gas concentrations and global climate in a substantial way for at least 6,000 years.

© 2008-2021. All rights reserved. Any copying, redistribution or retransmission of any of the contents of this service without the expressed written permission of the American Geosciences Institute is expressly prohibited. Click here for all copyright requests.