How the World Really Works: A Scientist’s Guide to Our Past, Present and Future

by Vaclav Smil

In 1872, a century after the appearance of the last volume of the Encyclopédie, any collection of knowledge had to resort to the superficial treatment of a rapidly expanding range of topics, and, one and a half centuries later, it is impossible to sum up our understanding even within narrowly circumscribed specialties: such terms as “physics” or “biology” are fairly meaningless labels, and experts in particle physics would find it very hard to understand even the first page of a new research paper in viral immunology. Obviously, this atomization of knowledge has not made any public decision-making easier. Highly specialized branches of modern science have become so arcane that many people employed in them are forced to train until their early or mid-thirties in order to join the new priesthood.

Urbanization and mechanization have been two important reasons for this comprehension deficit. Since the year 2007, more than half of humanity has lived in cities (more than 80 percent in all affluent countries), and unlike in the industrializing cities of the 19th and early 20th centuries, jobs in modern urban areas are largely in services. Most modern urbanites are thus disconnected not only from the ways we produce our food but also from the ways we build our machines and devices, and the growing mechanization of all productive activity means that only a very small share of the global population now engages in delivering civilization’s energy and the materials that comprise our modern world.

An average inhabitant of the Earth nowadays has at their disposal nearly 700 times more useful energy than their ancestors had at the beginning of the 19th century. Moreover, within a lifetime of people born just after the Second World War the rate had more than tripled, from about 10 to 34 GJ/capita between 1950 and 2020. Translating the last rate into more readily imaginable equivalents, it is as if an average Earthling has every year at their personal disposal about 800 kilograms (0.8 tons, or nearly six barrels) of crude oil, or about 1.5 tons of good bituminous coal. And when put in terms of physical labor, it is as if 60 adults would be working non-stop, day and night, for each average person; and for the inhabitants of affluent countries this equivalent of steadily laboring adults would be, depending on the specific country, mostly between 200 and 240. On average, humans now have unprecedented amounts of energy at their disposal.

Lubricants are needed to minimize friction in everything from the massive turbofan engines in wide-body jetliners to miniature bearings.38 Globally, the automotive sector, now with more than 1.4 billion vehicles on the road, is the largest consumer, followed by use in industry—with the largest markets being textiles, energy, chemicals, and food processing—and in ocean-going vessels. Annual use of these compounds now surpasses 120 megatons (for comparison, global output of all edible oils, from olive to soybean, is now about 200 megatons a year), and because the available alternatives—synthetic lubricants made from simpler, but still often oil-based, compounds rather than those derived directly from crude oil—are more expensive, this demand will grow further as these industries expand around the world. Another product derived from crude oil is asphalt. Global output of this black and sticky material is now on the order of 100 megatons, with 85 percent of it going to paving (hot and warm asphalt mixes) and most of the rest to roofing.39 And hydrocarbons have yet another indispensable non-fuel use: as feedstocks for many different chemical syntheses (dominated by ethane, propane, and butane from natural gas liquids) producing a variety of synthetic fibers, resins, adhesives, dyes, paints and coatings, detergents, and pesticides, all vital in myriad ways to our modern world.40 Given these advantages and benefits, it was predictable—indeed unavoidable—that our dependence on cru...

Global oil extraction of crude oil doubled during the 1950s, and by 1964 crude oil surpassed coal as the world’s most important fossil fuel, but although its output kept on rising, supply remained plentiful and so prices were falling. In constant (inflation-adjusted) monies, the world oil price was lower in 1950 than it was in 1940, lower in 1960 than in 1950—and lower still in 1970 than in 1960.43 Not surprisingly, demand was coming from all sectors. In real terms, crude oil was so cheap that there were no incentives to use it efficiently: American houses in regions with a cold climate, increasingly heated by oil furnaces, were built with single-glazed windows and without adequate wall insulation; the average efficiency of American cars actually declined between 1933 and 1973; and energy-intensive industries continued to operate by using inefficient processes.44 Perhaps most notably, America’s pace of replacing old open-hearth furnaces with superior oxygen furnaces to make steel was much slower than in Japan and Western Europe.

On January 1 1974, the Gulf states raised their posted price to $11.65/barrel, completing a 4.5-fold rise in the cost of this essential energy source in a single year—and this ended the era of rapid economic expansion that had been energized by cheap oil. From 1950 to 1973 the Western European economic product had nearly tripled, and the US GDP had more than doubled in that single generation. Between 1973 and 1975 the global economic growth rate dropped by about 90 percent, and as soon as the economies affected by higher oil prices began to adjust to these new realities—above all by impressive improvements in industrial energy efficiency—the fall of the Iranian monarchy and the takeover of Iran by a fundamentalist theocracy led to a second wave of oil price rises, from about $13 in 1978 to $34 in 1981, and to another 90 percent decline in the global rate of economic growth between 1979 and 1982.47 More than $30 a barrel was a demand-destroying price and by 1986 oil was again selling at just $13 a barrel, setting the stage for yet another round of globalization—this time centered on China, whose rapid modernization was driven by Deng Xiaoping’s economic reforms and by massive foreign investment. Two generations later, only those who lived through those years of price and supply turmoil (or those, increasingly few, who studied their impact) appreciate how traumatic these two waves of price rises were. Consequences of the resulting economic reversals are still felt four decad...

How soon will we fly intercontinentally on a wide-body jet powered by batteries? News headlines assure us that the future of flight is electric—touchingly ignoring the huge gap between the energy density of kerosene burned by turbofans and today’s best lithium-ion (Li-ion) batteries that would be on board these hypothetically electric planes. Turbofan engines powering jetliners burn fuel whose energy density is 46 megajoules per kilogram (that’s nearly 12,000 watt-hours per kilogram), converting chemical to thermal and kinetic energy—while today’s best Li-ion batteries supply less than 300 Wh/kg, more than a 40-fold difference.79 Admittedly, electric motors are roughly twice as efficient energy converters as gas turbines, and hence the effective density gap is “only” about 20-fold. But during the past 30 years the maximum energy density of batteries has roughly tripled, and even if we were to triple that again densities would still be well below 3,000 Wh/kg in 2050—falling far short of taking a wide-body plane from New York to Tokyo or from Paris to Singapore, something we have been doing daily for decades with kerosene-fueled Boeings and Airbuses.

When asked to give common examples of our reliance on fossil fuels, inhabitants of the colder parts of Europe and North America will think immediately about the natural gas used to heat their houses. People everywhere will point out the combustion of liquid fuels that power most of our transportation but the modern world’s most important—and fundamentally existential—dependence on fossil fuels is their direct and indirect use in the production of our food. Direct use includes fuels to power all field machinery (mostly tractors, combines, and other harvesters), the transportation of harvests from fields to storage and processing sites, and irrigation pumps. Indirect use is much broader, taking into account the fuels and electricity used to produce agricultural machinery, fertilizers, and agrochemicals (herbicides, insecticides, fungicides), and other inputs ranging from glass and plastic sheets for greenhouses, to global positioning devices that enable precision farming.

This is laborious, slow, and low-yielding farming—but it is completely solar, and no other energy inputs are required beyond the Sun’s radiation: the crops produce food for people and feed for animals; trees yield wood for cooking and heating; and wood is also used to make metallurgical charcoal for smelting iron ores and producing small metal objects including plow plates, sickles, scythes, knives, and strakes to cover wooden wagon wheels. In modern parlance, we would say that this farming requires no non-renewable (fossil fuel) energy inputs and only a minimum of non-renewable material subsidies (iron components, stones for gristmills), and that the production of both crops and materials relies solely on renewable energies deployed through the exertion of human and animal muscles.

This is a new, hybrid kind of farming, as the indispensable solar input is augmented by non-renewable anthropogenic energies derived overwhelmingly from coal. The new arrangement requires more animal labor than human labor, and as working horses (and mules in the American South) need grain feed—mainly oats—as well as fresh grass and hay, their large numbers make substantial demands on the country’s crop production: about one-quarter of all American farmland is devoted to growing fodder for draft animals.

In two centuries, the human labor to produce a kilogram of American wheat was reduced from 10 minutes to less than two seconds. This is how our modern world really works. And as mentioned, I could have done similarly stunning reconstructions of falling labor inputs, rising yields, and soaring productivity for Chinese or Indian rice. The time frames would be different but the relative gains would be similar.

But the energy required to make and to power farm machinery is dwarfed by the energy requirements of producing agrochemicals. Modern farming requires fungicides and insecticides to minimize crop losses, and herbicides to prevent weeds from competing for the available plant nutrients and water. All of these are highly energy-intensive products but they are applied in relatively small quantities (just fractions of a kilogram per hectare).14 In contrast, fertilizers that supply the three essential plant macronutrients—nitrogen, phosphorus, and potassium—require less energy per unit of the final product but are needed in large quantities to ensure high crop yields.

In preindustrial cropping, the wastes had to be collected in villages, towns, and cities, fermented in heaps or pits and—because of their low nitrogen content—applied to fields in massive amounts, commonly 10 tons per hectare but sometimes up to 30 tons (the latter mass being equivalent to 25–30 small European cars), in order to provide the needed nitrogen. Not surprisingly, this was commonly the most time-consuming task in traditional farming, claiming at least a fifth, and as much as a third, of all (human and animal) labor in cropping.

Given that vegans extol eating plants, and that the media have reported extensively on the high environmental cost of meat, you might think that gains in the energy cost of chicken have been surpassed by those in the cultivation and marketing of vegetables. You would be mistaken to think that. The opposite has been true, in fact, and there is no better example to illustrate these surprisingly high energy burdens than taking a close look at tomatoes.

This means that when bought in a Scandinavian supermarket, tomatoes from Almería’s heated plastic greenhouses have a stunningly high embedded production and transportation energy cost. Its total is equivalent to about 650 mL/kg, or more than five tablespoons (each containing 14.8 milliliters) of diesel fuel per medium-sized (125 gram) tomato! You can stage—easily and without any waste—a tabletop demonstration of this fossil fuel subsidy, by slicing a tomato of that size, spreading it out on a plate, and pouring over it 5–6 tablespoons of dark oil (sesame oil replicates the color well). When sufficiently impressed by the fossil fuel burden of this simple food, you can transfer the plate’s contents to a bowl, add two or three additional tomatoes, some soy sauce, salt, pepper, and sesame seeds, and enjoy a tasty tomato salad. How many vegans enjoying the salad are aware of its substantial fossil fuel pedigree?

If you want to eat wild fish with the lowest-possible fossil carbon footprint, stick to sardines. The mean for all seafood is stunningly high—700 mL/kg (nearly a full wine bottle of diesel fuel)—and the maxima for some wild shrimp and lobsters are, incredibly, more than 10 L/kg (and that includes a great deal of inedible shells!).43 This means that just two skewers of medium-sized wild shrimp (total weight of 100 grams) may require 0.5–1 liters of diesel fuel to catch—the equivalent of 2–4 cups of fuel.

Obviously, if all high-income countries were to follow these examples, they could reduce their crop harvests—because most of their grain harvests are not destined directly for food but for animal feed.75 But this is not a universal option. While meat intakes in many affluent countries have been declining and could be cut even further, they have been rising rapidly in such modernizing nations as Brazil and Indonesia (where they have more than doubled since 1980) and China (where they have quadrupled since 1980).76 Moreover, there are billions of people in Asia and Africa whose meat consumption remains minimal and whose health would benefit from more meaty diets.

These advances are, at present, a very long way off. They will depend on inexpensive renewable electricity generation backed up by adequate large-scale storage, a combination that is yet to be commercialized (and an alternative to large pumped hydro storage is yet to be invented; for more see chapter 3). A nearly perfect solution would be to develop grain or oil crops with the capabilities common to leguminous plants—that is, with their roots hosting bacteria able to convert inert atmospheric nitrogen to nitrates. Plant scientists have been dreaming about this for decades, but no releases of commercial nitrogen-fixing varieties of wheat or rice are coming anytime soon.80 Nor is it very likely that all affluent countries and better-off modernizing economies will adopt large-scale voluntary reductions in the quantity and variety of their typical diets, or that the resources (fuel, fertilizers, and machinery) saved by such pullbacks would be transferred to Africa to improve the continent’s still-dismal nutrition.

The best account of recent nitrogen flows in China’s agriculture shows that about 60 percent of the nutrient available to the country’s crops comes from synthetic ammonia: feeding three out of five of the Chinese population thus depends on the synthesis of this compound.23 The corresponding global mean is about 50 percent. This dependence easily justifies calling the Haber-Bosch synthesis of ammonia perhaps the most momentous technical advance in history. Other inventions, as William Crookes correctly judged, minister to our comforts, convenience, luxury, wealth, or productivity, and others yet save our lives from premature death and chronic disease—but without the synthesis of ammonia, we could not ensure the very survival of large shares of today’s and tomorrow’s population.

Ironmaking is highly energy-intensive, with about 75 percent of the total demand claimed by blast furnaces. Today’s best practices have a combined demand of just 17–20 gigajoules per ton of finished product; less efficient operations require 25–30 GJ/t.76 Obviously, the energy cost of secondary steel made in EAFs is much lower than the cost of integrated production: today’s best performance is just above 2 GJ/t. To this must be added the energy costs of rolling the metal (mostly 1.5–2 GJ/t), and hence the representative global rates for the overall energy cost may be about 25 GJ/t for integrated steelmaking and 5 GJ/t for recycled steel.77 The total energy requirement of global steel production in 2019 was about 34 exajoules, or about 6 percent of the world’s primary energy supply. Given the industry’s dependence on coking coal and natural gas, steelmaking has been also a major contributor to the anthropogenic generation of greenhouse gases. The World Steel Association puts the average global rate at 500 kilograms of carbon per ton, with recent primary steelmaking emitting about 900 megatons of carbon a year, or 7–9 percent of direct emissions from the global combustion of fossil fuels.

Between 1990 and 2020, the mass-scale concretization of the modern world has emplaced nearly 700 billion tons of hard but slowly crumbling material. The durability of concrete structures varies widely: while it is impossible to offer an average longevity figure, many will deteriorate badly after just two or three decades while others will do well for 60–100 years. This means that during the 21st century we will face unprecedented burdens of concrete deterioration, renewal, and removal (with, obviously, a particularly acute problem in China), as structures will have to be torn down—in order to be replaced or destroyed—or abandoned. Concrete structures can be slowly demolished, reinforcing steel can be separated, and both materials can be recycled: not cheap, but perfectly possible. After crushing and sieving, the aggregate can be incorporated in new concrete, and reinforcing steel can be recycled.101 Even now, replacement concrete and new concrete are needed everywhere.

If low labor costs were the sole reason for locating new factories abroad—as many people seem to erroneously believe—then sub-Saharan Africa would be the most obvious choice, and India would almost always be preferable to China. But during the second decade of the 21st century, China averaged about $230 billion of foreign direct investment a year, compared to less than $50 billion for India and just around $40 billion for all of sub-Saharan Africa (excluding South Africa).9 China provided a combination of other attractors—above all, centralized one-party government that could guarantee political stability and acceptable investment conditions; a large, highly homogeneous and literate population; and an enormous domestic market—that made it the preferred choice over Nigeria, Bangladesh, and even India, resulting in a remarkable collusion between the world’s largest communist state and a nearly complete lineup of the world’s leading capitalist enterprises.

In its most fundamental physical way, globalization is, and will remain, simply the movement of mass—of raw materials, foodstuffs, finished products, and people—and the transmission of information (warnings, guidance, news, data, ideas) and investment within and among the continents, enabled by techniques that make such transfers possible on large scales and in affordable and reliable ways.

The first steam-powered westward transatlantic crossings took place in 1838, but sailing ships remained competitive for another four decades. With wind as the prime mover, the cost of carrying a unit of cargo per unit of distance by a sailing ship was largely independent of the length of the voyage; while the longer the steamship voyage, the more of the vessel’s deadweight capacity had to be loaded with coal to fuel relatively inefficient engines, leaving less room for cargo. Refueling stations reduced, but did not eliminate, this disadvantage.27

This distinct and intensive, but still far from universal, spell of post-1950 globalization—which ended in 1973–1974 with OPEC’s two rounds of oil price increases and which was followed by 15 years of relative stagnation—was enabled by a combination of four fundamental technical advances. These were the rapid adoption of much more powerful and efficient designs of diesel engines; the introduction (and even faster diffusion) of a new prime mover, the gas turbine used for the propulsion of jetliners; superior designs for intercontinental shipping (massive bulk carriers for liquids and solids, and the containerization of other cargoes); and quantum leaps in computing and information processing.

The years between 1950 and 1973 were marked by rapid economic growth in virtually every part of the world: its global annual mean rate and its average per capita gains were nearly 2.5 times greater than during the previous globalization wave of 1850–1913, and the value of exported goods in the world economic product rose from a low of just over 4 percent in 1945 to 9.6 percent in 1950 and about 14 percent in 1974, equaling the 1913 share but with trade volume nearly ten times higher.65 Economic growth was nearly universal (China’s Great Famine years of 1958–1961 were the most consequential exception) but the benefits of this golden age of economic expansion—the postwar rebound, with high rates of growth helping to decrease economic inequality—were disproportionately concentrated in the West: by 1973, North America and the countries of Western Europe accounted for more than 60 percent of global exports.66 As the major Western European economies (Germany, UK, France) and Japan became the era’s most dynamic traders, inevitably America’s share in world trade gradually eroded.

As for maximum ship sizes, in 1972 and 1973 Malcolm McLean launched his largest container vessels, each with a capacity of 1,968 standard steel containers (nearly five times larger than his first converted vessels in 1957). In 1996, Regina Maersk could load 6,000 standard units; by 2008 the maximum was 13,800; and in 2019 the Mediterranean Shipping Company put into service six giant vessels each able to carry 23,756 standard containers, hence a 12-fold increase of maximum vessel capacity between 1973 and 2019.77 Inevitably, this mass-scale conversion to container shipping required the commensurate conversion of freight trains and truck transportation, and these intermodal chains now bring the goods from a city in China’s interior to the loading dock of a Walmart in Missouri.

The 105 mark (100,000 transistors) was reached in 1982, and in 1996, to celebrate the machine’s 50th anniversary, a group of students at the University of Pennsylvania recreated ENIAC by putting 174,569 transistors on a 7.4 mm × 5.3 mm silicon microchip: the original machine was more than 5 million times heavier, it required about 40,000 times more electricity, and the recreated chip was 500 times faster.83 And the progress continued: the 108 mark was surpassed in 2003, 109 in 2010, and by the end of 2019 AMD released its Epyc CPU with 39.5 billion transistors.84 This means that between 1971 and 2019 microprocessor power increased by seven orders of magnitude—17.1 billion times, to be exact.

This development was soon obscured by the Great Recession of 2008, but McKinsey’s analysis of 23 industry value chains (interconnected activities, from design to retail, that deliver final products) spanning 43 countries between 1995 to 2017 shows that goods-producing value chains (still growing slowly in absolute terms) have become significantly less trade-intensive, with exports declining from 28.1 percent of gross output in 2007 to 22.5 percent in 2017.96 What I see to be the study’s second-most important finding is that, contrary to common perception, only about 18 percent of the global goods trade is now driven by lower labor costs (labor arbitrage), that in many chains this share has been declining throughout the 2010s, and that global value chains are becoming more knowledge-intensive and rely increasingly on highly skilled labor. Similarly, an OECD study shows that the expansion of global value chains stopped in 2011 and since then has slightly declined: there has been less trade in intermediate goods and services.97

The latest published surveys show Japan and the US to be surprisingly close in total food energy consumed per day. In 2015–2016, US males consumed only 11 percent more, and US women not even 4 percent more food energy per day than their Japanese counterparts did in 2017. The two countries diverged moderately in total carbohydrate (Japan was ahead by less than 10 percent) and protein (with Americans less than 14 percent ahead) consumption, and both nations were well above the needed protein minima. But there is a major gap in terms of average fat intake, with American males consuming about 45 percent more and women 30 percent more than the Japanese. And the greatest disparity is in sugar intake: among US adults it is about 70 percent higher. When recalculated in terms of average annual differences, Americans have recently consumed about 8 kilograms more fat and 16 kilograms more sugar every year than the average adult in Japan.

In his pioneering 1969 analysis of risks, Chauncey Starr—at that time the dean of the School of Engineering and Applied Science at the University of California in Los Angeles—stressed the major difference in risk tolerance between voluntary and involuntary activities.20 When people think that they are in control (a perception that may be incorrect but that is based on previous experiences and hence on the belief that they can assess the likely outcome), they engage in activities—climbing vertical rock faces without ropes, skydiving, bullfighting—whose risks of serious injury or fatality may be a thousand-fold higher than the risk associated with such dreaded involuntary exposure as a terrorist attack in a large Western city. And most people have no problem engaging daily and repeatedly in activities that temporarily increase their risk by significant margins: hundreds of millions of people drive every day (and many apparently like to do so), and an even higher risk is tolerated by an even larger number of smokers21—in affluent countries, decades of education has reduced their ranks, but worldwide there are still more than 1 billion of them.

Perhaps the most stunning contrast of nuclear-related risk perceptions is seen when comparing France and Germany. France has been deriving more than 70 percent of its electricity from nuclear fission since the 1980s and nearly 60 reactors dot the country’s landscape, cooled by water from many French rivers, including the Seine, Rhine, Garonne, and Loire.25 Yet the longevity of the French population (second only to Spain within the EU) is the best testimony to the fact that these nuclear power plants have not been a discernible source of ill health or premature deaths—but across the Rhine it is not only the German Greens who believe that nuclear power is an infernal invention that must be eliminated as fast as possible, but much larger portions of society too.

Scheduled commercial flights, already a very low-risk activity at the end of the last century, got appreciably safer during the first two decades of the 21st century. This conclusion stands despite some disturbing recent losses, including the still-unsolved (and likely never to be explained) disappearance of Malaysia Airlines flight 370 somewhere over the Indian Ocean in March 2014, followed by the downing of Malaysia Airlines flight 17 over eastern Ukraine in July 2014, and the two crashes of the new Boeing 737 MAX—Lion Air flight 610 in the Java Sea (October 29, 2018) and Ethiopian Airlines flight 302 near Addis Ababa (March 10, 2019).53 Perhaps the most revealing way to compare the airline industry’s fatalities is per 100 billion passenger-kilometers flown. This rate was 14.3 in 2010, it reached a record low of 0.65 in 2017, but it increased to 2.75 in 2019. Flying in 2019 was thus more than five times safer than in 2010, and more than 200 times safer than at the beginning of the jetliner era in the late 1950s.54 Expressing these fatalities in terms of risks per hour of exposure is fairly straightforward. The mean 2015–2019 total of accidental deaths was 292; the averages of 68 trillion passenger-kilometers flown and 4.2 billion passengers mean that average passengers flew about 1,900 kilometers and spent about 2.5 hours in flight; the total of about 10.5 billion passenger-hours spent aloft and 292 fatalities translates to 2.8 × 10-8 (0.000000028) fatalities per person ...

As COVID-19 demonstrated yet again (and on scales that must have surprised even those who do not expect any good news), we are repeatedly caught ill-prepared for dealing with recurrent high-impact but relatively low-frequency risks such as viral pandemics that take place once in a decade, once in a generation, or once in a century. How would we then cope (all reports and analyses aside) with another Carrington Event, or with an asteroid hitting the ocean near the Azores and causing a massive circum-Atlantic tsunami of the same magnitude as the one caused by the 2011 Tōhoku earthquake—that is, up to 40 meters high and traveling up to 10 kilometers inland?

Given the continent’s still-high fertility rates, further expansion of cultivated land will be inevitable in Africa, but only limited extensions should take place in most of Asia, while Europe, North America, and Australia (with already-excessive food production and aging populations) should see further declines in cultivated land. The amount of land used in food production could be reduced with the combination of better farming practices, reduced food waste, and the widespread adoption of moderate meat consumption. As already explained in chapter 2, reversion to preindustrial farming is inconceivable in a world of nearly 8 billion people, but getting higher yields with existing inputs (agricultural intensification) conforms to a long-established trend, and the elimination of many wasteful practices could produce higher yields even with reduced fertilizer or pesticide use. A convincing large-scale, decade-long (2005–2015) demonstration of this included nearly 21 million farmers cultivating about a third of China’s cropland: they were able to raise staple grain yields by 11 percent while reducing the application of nitrogen per hectare by 15–18 percent.

As already explained, the worldwide efficiency of nitrogen uptake by crops has declined to less than 50 percent, and to below 40 percent in China and France. In conjunction with phosphorus, soluble nitrogen compounds contaminate waters and support excessive algal growth. Decomposing algae consume oxygen dissolved in seawater and create oxygen-less (anoxic) waters where fish and crustaceans cannot survive. These oxygen-depleted zones are prominent along the eastern and southern coasts of the United States and along coasts in Europe, China, and Japan.34 There are no easy, inexpensive, and rapid solutions to these environmental impacts. Better agronomic management (crop rotations, split applications of fertilizers to minimize their losses) is essential, and reduced meat consumption would be the single-most important adjustment as it would lower the need for producing feed grains—but sub-Saharan Africa will need much more nitrogen and phosphorus if it is to avoid chronic dependence on food imports.

The Earth’s atmosphere absorbs the incoming (short-wave) solar radiation and radiates (longer waves) to space. Without it, the temperature of the Earth would be -18oC, and hence our planet’s surface would be perpetually frozen. Trace gases change the planet’s radiation balance by absorbing some of the outgoing (infrared) radiation and raising the surface temperature. This allows for the existence of liquid water, whose evaporation puts water vapor (another gas that absorbs outgoing infrared, invisible waves) into the atmosphere. The overall result is that the Earth’s surface temperature is 33oC higher than it would be in the absence of trace gases and water vapor, and the global mean temperature of 15oC supports life in its many forms. Labeling this natural phenomenon as the “greenhouse effect” is a misleading analogy, because the heat inside a greenhouse is there not only because the glass enclosure prevents the escape of some infrared radiation but also because it cuts off air circulation. In contrast, the natural “greenhouse effect” is caused solely by the interception of a small share of outgoing infrared radiation by trace gases—while the global atmosphere remains in constant, unimpeded, and often violent motion. Water vapor is by far the most important absorber of outgoing radiation, and hence it is the gas that has been responsible for most of the atmospheric warming in the past and it will remain so in the future. Water vapor is the principal generator of the natur...

Nighttime temperatures are increasing faster than the daytime averages mainly because the boundary layer (the atmosphere just above the ground) is very thin—mere hundreds of meters—during the night, compared to several kilometers during the day, and hence it is more sensitive to warming.

The late-1980s “discovery” of carbon dioxide–induced global warming thus came more than a century after Foote and Tyndall made the link clear, nearly four generations after Arrhenius published a good quantitative estimate of the possible global warming effect, more than a generation after Revelle and Suess warned about an unprecedented and unrepeatable planet-wide geophysical experiment, and a decade after modern confirmation of climate sensitivity. Clearly, we did not have to wait for new computer models or for the establishment of an international bureaucracy to be aware of this change and to think about our responses.

But most studies concur that demand-driven freshwater scarcity will have a much greater impact than the shortages induced by climate change. As a result, our best option for dealing with future water supply is to manage demand, and one of the best large-scale examples of this working is the recent history of the US’s reducing per capita water usage.63 In 2015, overall US water consumption was less than 4 percent higher than it was in 1965—but during the intervening 50 years the country’s population increased by 68 percent, its GDP (in constant monies) more than quadrupled, and irrigated farmland increased by about 40 percent. This means that average per capita water use decreased by nearly 40 percent, that the water intensity of the US economy (units of water per unit of constant GDP) declined by 76 percent, and, as the total volume of water used for irrigation was actually slightly lower in 2015, that applications per unit of farmland declined by nearly a third. There are, of course, physical limits to the further reduction in all of these uses of water, but the US experience shows that the gains can go far beyond marginal ones.

A warmer atmosphere will also enhance water loss from plants (evapotranspiration), but that does not mean that crops and forests will wilt as they lose water. A rising atmospheric level of CO2 means that the water required per unit of yield will decrease in a warmer and CO2-richer biosphere. This effect has already been measured in some crops, and wheat and rice (staple grains which rely on the most common photosynthetic pathway) will increase their water use efficiency more than corn or sugar cane (which use a less common, but inherently more efficient, pathway).68 This means that, in some regions, wheat and other crops could yield as much or more than today, even if the precipitation they receive is reduced by 10–20 percent.

Similarly, we are now promised even more astonishing “disruptive” innovations and AI-driven “solutions.” The reality is that any sufficiently effective steps will be decidedly non-magical, gradual, and costly. We have been transforming the environment on increasing scales and with rising intensity for millennia, and we have derived many benefits from these changes—but, inevitably, the biosphere has suffered. There are ways to reduce those impacts but the resolve to deploy them at required scales has been lacking, and if we start acting in a sufficiently effective manner (and this now requires doing so on a global scale) we will have to pay a considerable economic and social price. Will we, eventually, do so deliberately, with foresight; will we act only when forced by deteriorating conditions; or will we fail to act in a meaningful way?

For generations, businesses and governments were the most common practitioners and consumers of forecasting, then academics joined the game in large numbers from the 1950s, and now anybody can be a forecaster—even without any mathematical skills—simply by using plug-in software or (as has been in vogue lately) by making baseless qualitative predictions. As in so many other cases of newly expanded endeavors (information flows, mass education), the quantity of modern forecasting has become inversely proportional to its quality. Many forecasts are nothing but the simplest extensions of past trajectories; others are the outcome of complicated interactive models which incorporate large numbers of variables and run using different assumptions each time (essentially the numerical equivalent of narrative scenarios); and some have hardly any quantitative component and are just wishful and exceedingly politically correct narratives.

The inertia of large, complex systems is due to their basic energetic and material demands—as well as the scale of their operations. Demands for energy and materials are constantly affected by the quest for higher efficiencies and for optimized production processes, but efficiency improvements and relative dematerialization have their physical limits, and advantages brought by new alternatives will have offsetting costs. Examples of such realities abound. Turning, once again, to two fundamental inputs, the theoretical minimum of primary energy needed to produce steel (combining blast furnace and basic oxygen furnace demands) is about 18 gigajoules per ton of hot metal, and ammonia cannot be synthesized from its elements with less than about 21 gigajoules per ton.32 One possible solution is to replace steel with aluminum. This lowers the mass of a specific design, but primary aluminum requires five to six times more energy to produce than primary steel, and it cannot be used in many applications that require steel’s much greater strength. The most radical way to cut energy costs and the environmental impact of nitrogen fertilizers is to reduce how much is used: that option is available to affluent countries with their excessive food supply and waste—but hundreds of millions of stunted children, mostly in Africa, need to drink more milk and eat more meat, and that protein can come only from substantially increasing the amount of nitrogen they use in cropping. Just to drive t...

Consequently, the US had to pay exorbitant prices to China—the country where the brilliant architects of globalization concentrated nearly all of the manufacturing of these essential items—in order to secure airlifts of inadequate amounts of protective equipment just to prevent hospital closures in the midst of a pandemic.44 The country that spends more than half a trillion every year on its military (more than all of its potential adversaries put together) was unprepared for an event that was absolutely certain to occur, and it did not have enough basic medical supplies: investment in domestic production worth a few hundred million dollars could have significantly reduced the economic losses of COVID-19, measured in the trillions!

Beyond that, we get into realms where it is easy to make order of magnitude mistakes: some rich families (business founders, owners, or lucky inheritors) now add annually to their holdings tens (107) or hundreds (108) of millions of dollars; in 2020 the world had about 2,100 billionaires (109 dollars), and the richest ones are now valued at more than $100 billion or 1011 dollars.2 In terms of individual net worth, compared with the few dollars’ worth of tattered clothes and the worn-out shoes of a destitute African migrant, the gap is thus 10 orders of magnitude. This difference is so large that we can find no equivalent among the properties separating the two most notable classes of terrestrial animals: birds and mammals. The difference between the body masses of the smallest and the largest mammals (the Etruscan shrew at 100 grams and the African elephant at 106 grams) is “just” six orders of magnitude. The difference between the wingspan of the smallest and the largest flying birds (the bee hummingbird at 3 centimeters and the Andean condor at 320 centimeters) is only two orders of magnitude.3 Clearly, some humans have gone far further than natural evolution ever could in separating themselves from the crowd.