Everywhere and Nowhere

Proxies​ stand in for something or someone else: a press secretary stands in for a politician; the number of citations a scientific paper gets stands in for its significance; the rise or fall of Gross Domestic Product stands in for the overall health of the economy. The proxy is never a perfect substitute for the person or thing it replaces. At best, the two are reliably correlated; at worst, only loosely and erratically so. Figuring out which proxies are generally dependable and under which circumstances is a complex business, and this is especially true of those sciences whose objects of inquiry are themselves proxies, such as sociology, economics and climatology. Entities such as society, the economy and climate are everywhere and nowhere. They are the product of many factors and must be observed indirectly via proxies. There’s a reason Adam Smith’s ‘invisible hand’ was invisible. This doesn’t make these entities any less real, but it does make studying them fiendishly complicated.

Climatology may be the ultimate proxy science. Weather is easy to measure and describe. It is local and short-term: whether it’s going to be rainy or sunny for the next few days; today’s temperature and barometer readings; tomorrow’s likely wind speed and direction. Climate is regional and middle to long-term: temperatures, humidity levels, the kinds and amount of precipitation, wind speeds and other factors averaged over decades, centuries, even millennia. Climate is understood today as a statistical entity comprising many different kinds of mostly quantitative data, harvested by standardised instruments positioned in an observation network. But at least two thousand years before the invention of thermometers and barometers, there was a concept of climate zones (the Greek klima originally meant ‘inclination’ or ‘latitude’, and later ‘region’). Aristotle, Hipparchus, Ptolemy, Strabo, Pliny the Elder and other ancient authorities carved up the Earth into horizontal bands (the number varied from five to seven), positing that the angle of the sun and the length of a day were the most important proxies for climate – hottest at the equator and coldest at the poles. Pliny’s Naturalis Historia drew inferences from climate zones to the complexion and character of the people who inhabited them: Ethiopians were scorched by the sun and made wise by their climate; northerners were made pale and fierce by theirs.

By happy coincidence, Pliny located the mean between these extremes in the Mediterranean temperate zone where he resided, claiming that he and his fellow Romans were ‘of medium bodily stature, with a marked blending even in the matter of complexion; customs are gentle, senses clear, intellect fertile and able to grasp the whole of nature.’ Although the early modern voyages of European commercial and imperial expansion eroded the credibility of Pliny’s geography and climate zones (imagine the disappointment of the English colonists who discovered that Massachusetts and Rome did not share the same climate), theories linking climate to national and racial character persisted well into the 19th century. If latitude was a proxy for climate, climate was a proxy for ethnic stereotypes.

The Russian-German scientist Wladimir Köppen created a new system of climate zones in 1884 on the basis of vegetation patterns, resulting in a patchwork of rainforests, deserts and tundra scattered across different latitudes and continents rather than in bands circling the globe. Yet there continued to be speculation about the way climate affected culture. The appeal of climate determinism isn’t hard to understand: what other factor has been so pervasive? As far back into the past as historical records stretched, the climates of the inhabited regions of the earth seemed relatively steady, despite the occasional drought or flood. The concept of climate as the mean of temperatures and precipitation over centuries meant that spiky fluctuations in the weather tended to be smoothed out. In the short run, contingency mattered, as did bouts of savage weather. But in the long run, climate and geography dominated.

Melissa Charenko’s book tracks what happened to climate science when new technologies vastly extended its chronological reach. In the mid-19th century, long-held assumptions about climate stability began to totter as geological evidence of past ice ages and the discovery of fossils of extinct species like the mastodon and megatherium emerged. Human archives, especially of past measurements made regularly on calibrated instruments, had the advantage of high temporal resolution and short chains of inference: if a well-kept record of thermometer and barometer readings for a given place was available, then the averages that linked weather to climate could be calculated. Some scientists, such as Harvard’s Robert DeCourcy Ward, who in 1910 became the first professor of climatology in the US, disdained all proxies, insisting that instrumental measurements were the only reliable indices for the modelling of climates past and present. But some believed they had discovered climate cycles, possibly correlated with sunspots, and others detected inexorable long-term trends. The Russian naturalist and anarchist Pyotr Kropotkin concluded from the dried lake beds he had seen in Siberia that the entire northern hemisphere was undergoing creeping desertification. Throughout the 19th and early 20th centuries, the balance of evidence swung back and forth between theories of climate stability and climate dynamism. There were too many indices from which to infer past climates, and they didn’t always align.

Instrument purists like Ward viewed all such indirect evidence with suspicion. But rigour came at a high price. Thermometers and barometers hadn’t been invented until the 17th century, and it wasn’t until the late 18th century that systematic measurements on standardised instruments were attempted – and even then only in a few parts of Europe and North America. Restricting the evidence for the reconstruction of past climates to calibrated instrument readings limited the geographical and temporal scope of climatology to a few hundred years. By contrast, nature’s archives of climate proxies elongated the timescale to many millennia and expanded its range to wherever scientific ingenuity could discover a proxy.

Proxies for past climates are various: fossilised plants and pollen preserved in peat bogs, the tree rings of giant sequoias, air bubbles trapped in ice cores, packrat middens. As in the case of geological strata, deeper usually means older. Tree rings and ice cores have an annual clock that can be reckoned backwards from now (the outermost ring of a newly felled tree or last winter’s layer of snow). These proxies offer the twin advantages of sharp temporal resolution and absolute dating, and in the case of some deep-bore ice cores they go back nearly 800,000 years. But they are limited to those regions with long-lived trees or permafrost. The past climates of other places must be inferred from more ambiguous proxies. Before radiometric dating became widely available in the mid-20th century, many proxies could only be dated relatively. The Scandinavian scientists who dug up pinecones and hazelnuts in peat bogs based their calculations on the depth at which such plant macrofossils were buried, much as archaeologists first dated successive settlements on the site of Troy by numbering them consecutively in order of excavation. Hazel trees flourish in temperate zones, and the Swedish geologist Gunnar Andersson inferred from the distribution of fossilised hazelnuts that parts of Sweden had once been significantly warmer (even after the last ice age) than they are today. But how long ago, exactly? Pollen proxies are more plentiful and versatile. As allergy sufferers know, trees can release immense amounts of pollen each spring, and the pollen grain’s hard protective walls decay very slowly. By identifying dominant plant genera in fossilised pollen excavated at different depths, scientists could match vegetation patterns with past climatic conditions. Determining just how long past was a trickier matter, as was ascertaining whether genera that happened to survive for thousands of years indicated the way ecological communities adapted to changes in climate.

Uncertainties plagued other proxies too. Take the smelly mounds of excrement deposited by the Shasta ground sloth, which once inhabited what is now the south-western United States. The sloths were herbivores and made their home in caves such as Rampart Cave near the Grand Canyon, where in the 1960s and 1970s the palaeoecologist Paul S. Martin began excavating layers of dung in the hope of discovering why the species had survived several glaciation periods, only to go extinct about ten thousand years ago. Martin thought the plant remains preserved in the sloth dung ruled out any sudden change of climate as the cause: judging from the fossil analysis, the sloths had been chomping on their usual diet up until the time they disappeared. He concluded that the sloths had been hunted to extinction by Paleoindians, who killed species not yet adapted to human predators well in excess of their need for food and fur. Known as the ‘overkill’ hypothesis, the idea of a will-to-kill was invoked by Martin and others to explain the disappearance of early Holocene megafauna after the arrival of modern humans in the Americas. But by the time Martin published his theory, radiocarbon dating could fix absolute dates and it showed a disparity of about three thousand years between the first appearance of humans in the area and the disappearance of the sloths, though fossilised plants in dung from other caves seemed to suggest a closer coincidence. Martin’s models predicted the extinction of megafauna within only a few hundred years of the arrival of humans.

It is impossible to avoid interpretation with proxy data, and the more proxies involved (and the more kinds of proxy), the greater the leeway for alternative interpretations. This is enough to fuel scientific controversy, even before politics are added to the mix. Charenko points out that the overkill hypothesis resonated with Cold War anxieties about the risk of engineering our own extinction through nuclear war and concomitant speculation among social scientists and sociobiologists that there might be an aggressive instinct in Homo sapiens. Indeed the term ‘overkill’ was borrowed from critics of the nuclear arms race, who argued that extant stockpiles of weapons were more than enough to wipe out the world’s population several times over. Particularly sensitive was the claim that it was the ancestors of Native Americans who had mowed down the ground sloths, woolly mammoths and other American fauna in a hunting spree. At a moment when environmentalists critical of industrial farming were finding common cause with Native American activists defending their rights to ancestral lands, Martin’s overkill hypothesis was seized on by defenders of the status quo in an effort to drive a wedge between the two groups. Oddly, everyone accepted the dubious identification of contemporary Native Americans with their ancestors of more than eleven thousand years ago, as if culture had stood still, neither diversifying nor developing, despite evidence to the contrary. Native American activists, for whom claims to cultural continuity were paramount, countered scientific accounts with rival oral traditions that allegedly reached back into the deep past.

Charenko is sympathetic to the Native American cause without endorsing the claims of oral traditions over those based on the evidence of scientific proxies. Maintaining such a position can be difficult: there is a fine line, for instance, between giving an honest account of the uncertainties and limitations involved in the use of scientific proxies and aiding climate change deniers who exploit uncertainty to sow doubt and stall action. As Naomi Oreskes and Erik Conway showed in Merchants of Doubt: How a Handful of Scientists Obscured the Truth on Issues from Tobacco Smoke to Global Warming (2010), a small group of scientists, employed first by tobacco companies and then by oil conglomerates, clouded the scientific consensus that smoking was bad for your health and that carbon dioxide emissions were heating up the atmosphere. Journalists became their unwitting accomplices because doctrines of ‘balance’, conflated with ‘objectivity’, disposed them to grant both sides equal airtime, even where the evidence and expert opinion overwhelmingly favoured one over the other. Charenko is forthright on the subject of global warming, which she considers conclusively proven, and is at pains not to abet the deniers. Yet the heart of her argument about scientific proxies is that they are intrinsically uncertain and that they have been used in the past to make extravagant claims later disproven.

Charenko’s dilemma – how to talk about uncertainty in science without undermining science – is not hers alone. Scientists have evolved refined methods for estimating the possible error associated with various kinds of evidence; indeed, being a credible scientist means being able to weigh evidence judiciously. Peer review is designed to assess the quality of the evidence, flag unwarranted inferences and measure the gap between evidence and interpretation. That gap can be narrowed but never closed: if it could be, science would not advance. Viewed in this light, uncertainty is a precondition for progress. But the matter is seldom presented in this way to the general public. Both scientists and journalists have motives for exaggerating scientific certainty and hyping results. Scientists, whose research is largely supported by public funds, have little to gain from being cautious and modest about their discoveries, especially since there are always rivals vying for the same few grants. For journalists, nothing ruins a headline faster than a qualifying ‘maybe’. The result of such overconfidence is the erosion of public trust in science. We badly need a grown-up way of talking about scientific uncertainty, uncontaminated by a long philosophical and theological tradition in which truths worthy of the name must be certain and eternal. Just because it is progressive, scientific knowledge will always fall short of that standard.

Charenko brings her history of climate proxies up to the present, in which the monstrously complex models made possible by supercomputers dominate projections of future climates. The multiproxy analyses used to reconstruct past climates also depend on the enormous increases in computing power since the 1990s. These computer-boosted methods represent a great advance in the ability of scientists to treat climate as a whole, rather than as a patchwork of different regional weather patterns studied by means of different proxies at different temporal resolutions. Yet it is just this holistic vision that makes Charenko uneasy. She prefers the local to the global, for reasons both practical and political. The analogue models of the second half of the 20th century, which paired present ecosystems with past ones, may have collapsed under the weight of their own detail, but they did succeed in showing that ecosystems don’t necessarily respond to climate changes as systems. Individual plant genera might or might not adapt, independent of their neighbours. What at a global level looks like overall devastation could at the local level reveal green shoots. Charenko takes this lesson one step further. Small-scale communities may offer more scope for agency and adaptation than the planet acting as a whole. She suggests that in navigating climate change ‘perhaps if we think smaller, we’ll act smarter.’ Just as she detects sympathetic vibrations between the overkill hypothesis of anthropogenic megafauna extinction and Cold War fears of imminent human extinction, there are undertones of the anti-globalisation politics of the current moment in her small-is-beautiful plea. Forget about the increasingly performative Conferences of the Parties orchestrated by the United Nations Framework Convention on Climate Change (COP31 is due to take place later this year); forget about the photograph of the cloud-streaked blue marble planet that first appeared on the cover of the Whole Earth Catalogue in 1968. For that matter, forget about the UN and the whole Earth altogether. Il faut cultiver notre jardin.

Belief in climate stability once fostered a belief in indelible ethnic types; belief in climate dynamism lent itself to deterministic explanations of any and all kinds of historical change. What all climate-based theories of history assumed until recently was the negligible role of human agency. Climate seemed to dwarf the puny efforts of human beings to change their lot. It was an unpleasant shock to discover that human beings could alter the climate, and not for the better. First came desertification on a regional scale, as in the case of the farmers who inadvertently turned the North American prairies into the Dust Bowl of the 1930s, and then warming on a global scale, as carbon dioxide emissions raised temperatures to levels that melted glaciers, unleashed floods and scorched cities. Fantasies of geoengineering schemes to reverse these effects only drive home the point that the roles are now reversed. Charenko writes that these revelations have challenged the function of climate proxies: what’s the use of reconstructing past climates if climate change has ‘ruptured’ the connection between past, present and future? The controlled uncertainty of the climate proxies, each carefully checked against the others, has given way to the radical uncertainty of a future in which all bets are off.