Version) This is the fully sourced and detailed version of Chapter 2: Too Many People. If you have no interest in these details, please skip forward to Chapter 3: Slaves to Growth on page 84.
This book is not designed for the reader to have to read every chapter. Only engage with the chapters that answer questions that you have. If at any point while reading the below you feel like you get it, stop reading. Move on to the next chapter. Diving deep into our current planetary reality is a lot. I suggest taking breaks when the evidence becomes too painful to hold.
First, I will start from the bottom and ask what clearly sustainable human life looks like. This gives us a baseline: a population range the Earth can support with the least guesswork. It also shows where the limits begin: food, water, soil, energy, habitat, and the living systems that must remain intact.
Second, I will show that today’s population of more than eight billion people is being supported by drawdown. The major examples are fresh water, topsoil, pollinators, fossil fuels, phosphorus, fixed nitrogen, ocean life, wildlife, forests, and wild habitat. I will focus on the cases where the drawdown is obvious and profound. The
goal is not to list every injury humans are causing; the goal is to show that the present system is not living from renewal.
I will not center climate change here. Climate change matters, but the problem is larger than climate. Even without climate change, a civilization that empties aquifers, erodes soil, clears habitat, exhausts fisheries, depends on fossil fuels, mines nutrients, and destroys wild populations is not sustainable.
Third, I will walk through the research that tries to estimate Earth’s human carrying capacity. The research is surprisingly limited, and still useful for our goals. The learnings come from seeing what the studies don’t show more than what they do show.
A paper that shows how much food modern agriculture can produce has not shown sustainable carrying capacity. It has shown that it is possible to have a lot of people when you draw down the Earth’s long-stored resources. A paper that assumes today’s crop yields while ignoring soil loss, groundwater depletion, fossil fuels, mined phosphorus, synthetic nitrogen, pesticides, and habitat destruction has not shown sustainable carrying capacity. A paper that assumes even higher future yields without explaining how that somehow becomes sustainable is fantasy. It has to explain how humans increase production while ending the drawdown that already supports the present system. A model that produces a carrying capacity number while ignoring multiple sustainability limits is
producing, at best, a theoretical limit for that subset of inputs—not a carrying capacity.
Fourth, I will look at the examples people often use to claim that humans have built sustainable societies that could be scaled. The test will be the same each time. Did the society live from renewal, or did it draw down principal?
Did it maintain the soil, water, forests, animals, fisheries, and habitat it depended on? Or did it expand, import resources, exhaust surrounding land, rely on conquest, or hide its damage somewhere else?
Fifth, I will explore building a high-end estimate while staying within this sustainability standard. I will give humans credit for skill, cooperation, restraint, ecological knowledge, and better technology. But I will not count temporary extraction as carrying capacity.
Let’s get to it.
What Does Sustainable Look Like?
About 12,000 years ago, before agriculture, the global human population was likely only a few million [1]. That slowly growing number was not the absolute maximum the Earth could sustainably hold. There were many places that humans could live where they did not yet live. Prehistoric humans did not all live sustainably, not even close. Well before destructive farming practices appeared, humans were
altering their environments to feed themselves. Humans mastered hunting quickly, and the historical record shows it.
When people showed up in new places, large animals often disappeared. A single woolly mammoth could feed a tribe of ten for months. But humans did not practice enough restraint; they killed these animals faster than the herds could replace themselves. They spent the biological principal, and the animals went extinct [2].
Over the hundreds of thousands of years that modern humans have existed, many groups did learn the harsh math of their environments. They realized through starvation that a river emptied of spawning fish would not feed their children. They understood that a plant taken without regard for regrowth would disappear. They hunted, gathered, fished, burned, tended, stored, and harvested in ways that took only what the Earth could replace. They learned to live off the interest.
These people actively managed the landscapes that fed them. They set fires at the right time, tended useful plants, and managed fish runs, orchards, shellfish beds, and animal habitats [3]. With this kind of knowledge, groups like the indigenous populations of the Pacific Northwest managed to survive off salmon runs for 10,000 years without depleting the resource [4]. Our lowest-case estimate for how many people the Earth can sustainably feed is not an estimate of unmanaged wilderness. It demands active human management and restraint.
The modern Earth is damaged. The salmon runs are a fraction of what they were. Wetlands have been drained, shellfish beds dredged, rivers dammed, and forests cut. Remaining sustainable tribes have been pushed to the least desirable lands by colonization and market pressure. So we can’t rely only on modern sustainable population densities. Instead, we take the physical size of Earth’s different ecosystems today and apply the highest historical population densities proven to work sustainably in those environments [5].
Start with the scientifically established baseline for habitable land—meaning land not covered by glaciers, permanent ice, or barren deserts. This 104 million square kilometers (km²) includes everything from wetlands to scrublands [6]. We can divide this surface into four physical categories and let the environmental limits do the math.
First, the high-productivity aquatic zones. The most energy-dense places on Earth for non-agricultural humans are coastal margins, estuaries, major river floodplains, and wetlands. These highly concentrated, resource-rich corridors cover roughly 5 million km² [7].
Assigning them the highest recorded ethnographic density—2 to 10 people per km²—supports 10 to 50 million people [8].
Second, the productive forests. Humans cannot eat pine needles, so if we subtract the frozen boreal forests and the dense interiors of deep rainforests, we are left with the world’s highly productive food mosaics: oak woodlands,
chestnut forests, hickory, palms, and rich river zones. This leaves roughly 25 million km² [9]. At 1 to 3 people per km², these systems support 25 to 75 million people [10].
Third, the open landscapes. Savannas, grasslands, and open woodlands cover roughly 30 million km² [11].
These biomes support people through scattered plants, seeds, roots, and game. But because energy is lost moving up the food chain to large animals, these meat-heavy seasonal systems support lower densities. At 0.25 to 1 person per km², they support 7.5 to 30 million people [12].
Finally, the marginal lands. Tundra, deep arid deserts, scrublands, and massive boreal pine forests account for the remaining 44 million km² [13]. Because of freezing temperatures, extreme heat, or low biological productivity, these regions support much thinner populations. At 0.02 to 0.3 people per km², this massive expanse supports just 0.8 to 13.2 million people [14].
If we lay these empirically grounded geographic areas alongside their corresponding ethnographic densities, the baseline math of a sustainably managed planet becomes visible:
Habitat Category Area Density Population Assumption Supported High-productivity 5M 2–10 10M–50M aquatic (coastal / km² people/km² wetland) Productive forest 25M 1–3 25M–75M (mixed woodland / km² people/km² riverine) Open landscapes 30M 0.25–1 7.5M–30 (savanna / grassland / km² people/km² M open woodland) Marginal lands (boreal 44M 0.02–0.3 0.8M–13.2 / tundra / arid) km² people/km² M Total 104M 43.3M–16 km² 8.2M That is our first anchor: roughly 40–170 million people. The low end of this figure is close to what we have actually seen on Earth: about 40 million people living purely on the Earth’s interest.
Draining Our Savings The gap between a range of 40–170 million people and our current 8.3 billion population is vast. How do we bridge that gap? We drain our ecological savings account.
It is common to view modern farming as a permanent expansion of the Earth’s carrying capacity. The assumption is that we learned how to make the pond bigger and breed fish faster. But looking closely at the physical mechanics of the global food system reveals a different picture. We have not fundamentally increased the Earth’s natural interest rate. Instead, the high crop yields of the last century are the result of drawing down stored resources. We can see this clearly by looking at the un-bypassable physical requirements for human expansion and agriculture: wild habitat, freshwater, soil, nutrients, and energy. In each category, the modern world relies on spending the principal.
First, consider wild habitat and the web of life. This is the Earth’s oldest savings account. Before we pumped the water or mined the soil, we expanded into forests, prairies, and wetlands where biological matter had been building up and cycling nutrients for over a billion years. We did not build our agricultural system in empty space; we exterminated the existing ecosystems to make room. It is sometimes estimated that roughly 20 to 23 percent of the Earth’s landmass remains as undisturbed “wilderness” [15].
However, that figure is misleading. It generally measures land free from heavy industrial infrastructure, and the vast
majority of that remaining space consists of uninhabitable land: deep deserts in North Africa and Australia, frozen tundra, and the dense boreal ice forests of the far north.
When we look for ecologically intact habitats—areas that still retain all their native animal species and have suffered zero human-caused extinctions—we find just 2.8 to 3 percent of the Earth’s land surface fits the bill [16]. In other words, humanity has altered, degraded, or destroyed roughly 97 percent of the planet’s terrestrial ecosystems, specifically monopolizing the zones most capable of supporting complex habitats. We replaced a living planet with shopping malls, parking lots, wheat farms, rice paddies, military bases, museums, factories, mines, poultry farms, pine tree plantations, oil refineries, sports stadiums, and pig slaughterhouses.
Since the dawn of human expansion, we have eradicated roughly half the mass of all plant life on Earth [17]. We have wiped out approximately 85 percent of the total biomass of wild land mammals [18].
The destruction does not stop at the shoreline.
Industrial fishing has stripped the oceans, eliminating roughly 90 percent of all large predatory fish, while whaling and hunting wiped out 80 percent of the total biomass of marine mammals [19]. In our rivers and lakes, freshwater vertebrate populations have collapsed by over 80 percent just in the last few decades [20]. In the skies, billions of wild birds have vanished.
Perhaps most dangerously, we are eradicating the foundation of the food web. As we clear land and apply global pesticides, insect populations are plummeting. In many studied areas, flying insect biomass has dropped by 75 to 80 percent over the last few decades alone [21].
Amidst this mass extinction, companies like Google are working with governments to find ways to use AI and robotics to drastically reduce wild mosquito populations by injecting factory-grown mosquitoes with diseases that make wild offspring unviable. At scale, this project would require releasing at least billions of biologically modified mosquitoes into the world daily, permanently. The project only works by releasing 100 times or more mosquitoes than the wild population it attacks. Scientists estimate that there are over 110 trillion mosquitoes alive at any point in time [22, 23].
For some reason, Google searches on this topic suggest that the plan is just for a single, rare mosquito type, while EPA filings have already been made for rolling out this technology on the most common mosquito types [24]. It also ignores that mosquitoes crossbreed. It is being pitched as a safe project, while something at this scale has never been attempted.
Mosquitoes are a foundational species for numerous ecosystems. Mosquito larvae live in still water, eat algae, bacteria, and rotting organic matter, and turn that material into living insect biomass that fish, amphibians, aquatic insects, turtles, birds, spiders, bats, and other
predators can eat. Adult mosquitoes also feed mostly on nectar, so are major pollinators.
The animals and ecosystems that use mosquitoes include pond, marsh, swamp, bog, wetland, floodplain, rice-field, tundra-pool, forest-pool, and stream-edge food webs; mosquitofish, killifish, minnows, guppies, sticklebacks, sunfish, bluegill, catfish, trout fry, salmon fry, dragonfly nymphs, damselfly nymphs, whirligig beetles, predaceous diving beetles, backswimmers, water boatmen, water striders, hydra, flatworms, spiders, robber flies, dance flies, frogs, toads, salamander larvae, newts, turtles, ducks, swallows, swifts, martins, flycatchers, warblers, nighthawks, bats, Arctic insect-eating birds, and more [25].
Because the Earth is a connected living system, extinctions cascade. When the insects die, the animals that eat them starve. Because these ecological unravelings take time to play out, we are carrying a massive “extinction debt.” We are currently in the Sixth Mass Extinction, and we will be dealing with the consequences of the life we have already destroyed for generations [26]. Still, we are actively cutting down the last tropical rainforests of the Amazon and the Congo basin.
Second, consider freshwater, our most immediate and unforgiving constraint. Rain is a renewable resource. If a society grows crops using only the rain that falls each year, or the natural flow of rivers fed by seasonal snowmelt, it is living within the system’s limits.
But rain alone cannot support global crop yields.
Much of the world’s agriculture relies on irrigation drawn from deep underground aquifers. These are ancient reserves, known as “fossil water,” that took tens of thousands of years to accumulate. The Ogallala Aquifer in the United States, which waters a massive portion of North American grain, took roughly 15,000 years to fill. Pumping has severely depleted parts of it in just a few decades, and once it is drawn down, it will take millennia to recharge [27].
The exact same process is playing out globally.
Satellite data shows groundwater depleting rapidly across the world’s major breadbaskets, including the North China Plain, the Central Valley of California, and northern India [28]. This is not a theoretical problem for the future. We are literally running out of water right now.
In 2018, Cape Town, South Africa, famously braced for “Day Zero,” the day the municipal water supply would be shut off entirely. In 2019, Chennai, India—a city of 11 million people—virtually ran out of groundwater, requiring trains to ship in drinking water. Today, massive urban centers like Mexico City face severe rationing as reservoirs and aquifers fail [29]. We are pumping ancient water to artificially inflate today’s carrying capacity.
Third, consider topsoil. Soil is a complex, living system that takes an incredibly long time to form. Under natural conditions, it takes roughly 500 to 1,000 years for the Earth to build a single inch of topsoil [30]. That slow accumulation is the interest.
Modern farming requires frequent plowing, monocropping, and leaving fields bare between harvests, exposing the soil to wind and rain. A stark historical example of what happens when we ignore this reality is the American Dust Bowl of the 1930s. People often conflate the devastating poverty of that era with the 1929 stock market crash, but the absolute destruction of the American Midwest was an ecological collapse. Farmers had spent decades deep-plowing the native, deep-rooted prairie grasses to plant endless fields of wheat. When a natural drought hit, there was no root system left to hold the earth in place.
Hundreds of millions of tons of topsoil simply blew away [31].
Today, agricultural land is eroding tens to hundreds of times faster than new soil can form [30]. The United Nations estimates that a third of the world’s soils are already moderately to highly degraded. Every year, millions of hectares of previously productive land degrade into desert-like conditions, a process known as desertification [32]. Farmers can temporarily mask the loss of healthy soil by adding chemical fertilizers, but those harvests are subsidized by washing centuries of biological wealth into the ocean.
Fourth, consider the nutrients that make plants grow. As intensive farming strips the soil, we have to artificially replace a broad list of minerals, including potassium, calcium, phosphorus, and nitrogen, just to keep yields from collapsing.
Like water, these must be physically extracted.
Potassium and agricultural lime are mined from the earth.
Phosphorus is mined from finite deposits of rock phosphate, located primarily in Morocco, China, and the United States. It took millions of years of geological pressure to create these reserves, and there is no synthetic substitute.
We blast it out of the ground, process it, and spread it on fields, where much of it washes away. High-grade phosphorus reserves are a strictly finite resource being drawn down globally [33].
Nitrogen tells a similar story. Before the 20th century, the amount of food humans could grow was limited by the natural nitrogen cycle. In the early 1900s, the Haber-Bosch process allowed humans to synthesize fertilizer by pulling nitrogen directly from the air. Today, synthetic nitrogen fertilizer feeds roughly half the human population [34].
But because this chemical process requires intense heat and pressure, it is powered almost entirely by fossil fuels, mostly natural gas and coal. The nitrogen in the food we eat, and therefore in the cells of our own bodies, is largely tied to fossil fuel extraction. New technologies, like “green ammonia,” attempt to make nitrogen fertilizer using renewable electricity and water instead of natural gas. Doing this on a global scale requires large amounts of fresh water and electricity, which shifts the drawdown elsewhere—requiring new infrastructure, land use, and the mining of metals for grids and batteries [35]. Furthermore,
applying massive quantities of nitrogen overloads ecosystems. The excess nitrogen washes off fields into rivers, creating massive dead zones in the oceans [36].
Finally, consider fossil fuels themselves: the engine of the entire system. Fossil fuels are the invisible muscle of modern agriculture. They power the tractors, mine the phosphorus, pump the water from deep aquifers, and manufacture the nitrogen. Most of the coal, oil, and natural gas we use comes from a massive biological anomaly hundreds of millions of years ago. Dense forests grew and died, but the biological world had not yet evolved the organisms capable of completely breaking down wood. The trees stacked up, were buried, and compressed over eons into dense carbon. It is a literal one-time geological deposit of ancient sunlight. It is not renewable.
Our entire civilization, and the food system that keeps eight billion people alive, is dependent on burning this one-time deposit. Even if we ignore the devastating impact burning fossil fuels has on the global climate, the physical reserves themselves are limited. Best estimates suggest that at our current rate of consumption, our proven, easily accessible reserves of oil and gas will be functionally exhausted in less than 100 years [37].
We did not cover everything in this section. We have not detailed resource depletion catastrophes such as decreased river flows, sinking cities on the coast that house 20% of humanity, rising sea levels, rising temperatures, air pollution, toxic waste and the poisoning of global water
systems, plastics, landfills, and more. We stop here not because we have cataloged every injury, but because the point has been proven enough.
If a household is living off a shrinking inheritance, they can eat lavishly for a long time. But the abundance on the table comes from emptying the savings account. That is what global agricultural drawdown looks like. And it sets the stage for the next problem: when researchers try to estimate how many people the Earth can hold, they usually assume this drawdown can continue indefinitely.
What Do Researchers Say Is Our Sustainable Limit?
You might expect researchers to have answered this question clearly by now. How many people can the Earth sustainably support?
They have not.
A major review looked at 69 studies that tried to estimate the limit to world population. The answers ranged from 0.5 billion to over a trillion people [38]. At first, that range sounds strange. But it makes sense once you see what the papers are doing. Most ecological researchers understand what sustainability means. They know a population cannot keep spending down soil, water, forests, fish, fossil fuels,
minerals, and wild habitat forever. They know carrying capacity, by definition, has to be based on renewal.
The carrying capacity literature contains only a handful of papers that even attempt to calculate a carrying capacity figure. The vast majority do not try; instead they attempt to calculate the maximum number of people that the Earth could feed for some period of time, often assuming we only look at a few variables.
The few examples that do attempt to produce a carrying capacity number all make drastic oversimplifications to how the modern world works, ignoring several components that would break if included in their models. Modern food systems are not just seeds and sunlight. They also depend on fertilizer, rain, irrigation, machinery, mined minerals, intensive organized labor, fossil fuels, pesticides, genetically modified seeds, feedlots, transportation, storage, refrigeration, roads, ports, power grids, global trade, and more. Remove those supports, and modern yields collapse. Keep those supports, and the system keeps drawing down the Earth’s principal.
That is the trap.
A paper can say we need renewable energy, better farming, less waste, nutrient recycling, soil protection, forest protection, and lower consumption. But saying what would be required is not the same as showing that it works. To prove sustainable carrying capacity, the paper would have to
show the replacement system actually feeding people without draining aquifers, losing soil, mining phosphorus, depending on fossil fuels, poisoning ecosystems, clearing habitat, or hiding the damage somewhere else. And that is just the food side of the equation. A sustainable system would also have to account for whatever amenities and luxuries beyond food that people would have in that model.
The literature does not do that.
The papers adequately show that the current population is too high. Then they offer a lower number that assumes humans can reorganize modern life before the damage catches up with us. That is useful. But it is not proof. The lowest estimate in the major review was 0.5 to 1.2 billion people [39]. At that lower and technically more defensible number, the paper still does not show the machinery of sustainability. It gives a number, not a demonstrated system. Other low estimates are more developed, but they follow the same pattern. Some estimate 1 billion. Some estimate 1.5 to 2 billion. Some estimate 2 to 3 billion. These are serious attempts. They are much more honest than papers that pretend modern production can continue forever. But they still tend to make the same leap: they name the conditions required for sustainability, then assume those conditions can be met at scale [40].
That is a very large assumption.
This should not surprise us. We already built the low-end estimate from the bottom up. When we looked at habitat, density, and proven renewal, the most defensible baseline was not billions. It was roughly 40 to 170 million people. For an academic paper to land there would mean saying something almost no modern institution is built to hear: that the sustainable human population may be closer to tens of millions than to billions. That is not a small correction to modern society. It is a rejection of the world’s basic growth story. The burden of proof remains where it belongs.
Anyone who claims the Earth can sustainably support billions of people has to show the system that does it. So far, that proof has not been given. Current models skip the hard part, rely on assumptions, and insert phrases like “renewable,” “circular,” or “efficient.” This preserves the hope that modern civilization can be tweaked into sustainability.
In defense of the literature, calculating the carrying capacity for humans is far more difficult than for most species because of our ability to use up resources in so many different ways. Current food yields far exceed what was being produced a century ago. Even after stripping the ocean of 90 to 95% of its large marine biomass in the past century, we are still taking far more food out of the ocean today than we were back then. We can take entire ecosystems and turn them into miles of grain. Other species do not do that.
When they use up their primary food source, they hit a limit
and die back. Humans expand. That skill masks the reality of our carrying capacity; unfortunately, it does not eliminate it.
The Earth’s reserves were massive, and some of them still are. While our sustainable carrying capacity could be ~170 million people, there is no doubt that a human population much larger than this can exist in a state of drawdown for a long time. We have done it. We are doing it.
The gap between our carrying capacity and what we can temporarily sustain with drawdown boggles the mind because our ability to consume entire habitats, capture and forcibly breed other animals, hunt, eliminate competitors, and adapt is otherworldly.
Our researchers are rarely asking what is sustainable. They are asking what else could we draw down to keep our population existing in overshoot.
Agriculture’s Finest The Aztec Chinampas Consider the wetland chinampas of the Basin of Mexico.
Modern agroecology literature frequently highlights this system as one of the most productive agricultural techniques ever developed [41]. People constructed these raised platforms in shallow lakes using dredged mud and organic silt. Water rose through the beds via capillary action, mitigating drought risks, while canal muck provided a
permanent, internal nutrient loop that could yield up to seven harvests per year.
The caloric output per hectare was exceptionally high. In the 1950s, the few remaining chinampas produced maize yields of 3.5 to 6.3 tons per hectare. Looking strictly at the crop footprint, estimates suggest that a single hectare of chinampas bed can produce enough food to support 15 to 20 people—which people often mistakenly interpret as supporting a population density of 1,500 to 2,000 people per square kilometer [42].
Calculating population density by looking only at the crop beds provides a distorted picture of the system’s carrying capacity. To understand how many people the landscape actually supported, you have to account for the entire physical and ecological footprint required to keep the system running.
First, the beds themselves only made up a fraction of the agricultural grid; the navigable canals separating them took up to half the surface area of the lakes. Beyond the water, the civilization required land for housing, plazas, roads, and waste. More significantly, the system required a massive, continuous supply of timber and fuel. Maize and beans must be boiled to be digestible, a process that then required immense amounts of firewood or charcoal.
Maintaining the chinampas walls required millions of willow stakes or ahuejote harvested from the surrounding environment. While the wetlands produced the food, the society required thousands of square kilometers of
surrounding dry hillsides and forests to supply cooking fuel, structural timber, and agave, or maguey, for clothing fibers.
Factoring in the land required for fuel, fiber, and infrastructure drastically expands the footprint per person.
Second, the system was maintained by immense physical exertion and a rigidly stratified social hierarchy. The earth was moved, the canals dredged, and the harvests transported entirely by human muscle using woven baskets, stone blades, and wooden digging sticks. The daily farming was done by the macehualtin (the common working class) and the mayeque (laborers legally bound to the estates of the nobility). Both groups were subjected to heavy tribute, meaning a large percentage of the food they grew was taken by the state. The macro-infrastructure that made the system possible—the dikes and aqueducts—was built using coatequitl. This was an unpaid, mandatory community labor draft where common people were forced to work.
Managers of the work whipped workers for motivation, and refusal to work resulted in arrest, beatings, being dragged to work, or being killed. Commoners were not free to leave the region or to stop working.
The chinampas were developed in the southern freshwater lakes of Xochimilco and Chalco centuries before the Aztec Empire existed. During this pre-imperial period (roughly 1150 to 1428 CE), archaeological surveys show the population of the entire 7,000-square-kilometer Basin of Mexico fluctuated between 200,000 and 300,000 people [43]. Factoring in the lakes, the forests, and the agricultural
zones, this gives a basin-wide density of roughly 30 to 45 people per km². This fairly self-sufficient system sustained a population of this size for nearly three centuries. This is a far greater population density than any people achieved from collecting nature’s interest alone.
Local Nahua sources do not describe this era as a flawless ecological balance. The Annals of Cuauhtitlan and the Codex Chimalpopoca document severe vulnerabilities [44]. While the lake water protected against minor frosts, the system was highly vulnerable to prolonged climate shocks.
In addition, Lake Texcoco—the lowest lake in the basin—was naturally brackish. During heavy rains, toxic saltwater would flood into the freshwater chinampas, poisoning the soil. The system survived only because of state-built infrastructure, like the 16-kilometer Nezahualcoyotl dike, erected specifically to keep the salt away from the crops [45].
The carrying capacity often associated with the chinampas centers on the 90-year imperial spike (1428–1521 CE) following the formation of the Aztec Triple Alliance. The system grew substantially, with far more volatility. The Great Famine of 1450–1454 devastated the people. Consecutive years of severe drought lowered the lake levels so far that capillary action failed, while early snows destroyed the maize. The system broke down.
Records show, “This was a time when they bought people…
At this time one sold oneself. One ate oneself; one
swallowed oneself. Or else one sold and delivered into bondage his beloved son, his dear child” [46].
More people means less buffer against hard times.
For the first few years of the drought, Moctezuma did what an Aztec ruler was expected to do: he opened the royal granaries and distributed the state’s massive, collected tribute reserves of maize to the public. However, by 1453 and 1454, the drought was still raging, and the state’s reserves were completely empty. He announced that the state would no longer arrest commoners for leaving the region, as their survival now demanded exit.
Thousands of starving people traveled to the Gulf Coast (the Totonac and Huastec regions), which had not been hit by the drought. The Totonac and Huastec people on the Gulf Coast lived in a lush, lower-altitude tropical region that had escaped the frost and drought. They had massive surpluses of corn. When the starving Nahua refugees arrived, the local Totonac merchants took advantage of the crisis, demanding human labor in exchange for food. The Nahua families either entered debt-bondage to keep themselves and their children from starving, or died.
The going rate was roughly 400 ears of corn for a young woman, and 500 for a young man [46].
The drought did eventually end, and the empire recovered.
By the time of the Spanish conquest, the population of the basin had reached roughly 1 million people [47]. The capital of Tenochtitlan was no longer feeding itself. The Codex Mendoza tribute rolls document that the city extracted thousands of tons of maize, beans, chia, amaranth, cotton, and firewood annually from a conquered empire. That empire covered roughly 300,000 square kilometers and contained an estimated 5 to 6 million people [48]. When you distribute the total population across the full territorial footprint required to sustain the empire and subsidize the capital, the societal density drops to roughly 16 to 20 people per km². Peak urban density in the basin required a massive rural extraction zone.
Stripped of empire and external tribute, demographic estimates of the basin’s theoretical maximum self-sufficient carrying capacity place the ceiling at 400,000 to 500,000 people [43]. This limit supports 55 to 70 people per km². Again, agricultural populations at their limits are highly vulnerable to weather and pest shocks.
Still, this is a very dense population. It is 5–7 times greater than the highest density used in the sustainable model shared at the beginning of this chapter. The chinampas system is limited by geography and could not be applied everywhere. It requires shallow, slow-moving freshwater wetlands, tropical or temperate climates where lakes do not freeze solid, and a massive local supply of
organic material and timber. It could reasonably be adapted to places like the Tonle Sap basin in Cambodia, the edges of the Mekong Delta, or the shallow highland lakes of the Andes, where a similar ancient raised-bed system called waru waru was developed [49].
Highly productive, shallow freshwater wetlands make up less than 1% of the Earth’s habitable surface [50].
Ultimately, the chinampas were physically destroyed. Following the conquest in 1521, the Spanish did not understand or value the wetland ecology. To build a European-style capital and prevent seasonal flooding, they aggressively drained the lakes. Without the water, the thermal mass vanished, the soil dried and sank, and the basin became plagued by dust storms [51].
Bali’s Subak Bali’s subak irrigation network is frequently held up as a textbook example of sustainable, cooperative agriculture.
When UNESCO designated the subak a World Heritage site, it praised the system’s ability to support a dense population on a rugged volcanic island. The official inscription highlighted a network of protected forests, terraces, canals, and water temples that have managed the ecology of entire watersheds since the eleventh century [52].
Modern writers often present the subak as a model of long-term ecological harmony, pointing out that these terraces produced grain for over a thousand years without a
drop in yield, managed by local farmers without centralized control.
This reputation dates back to 1597, when a scurvy-ridden Dutch expedition arrived in Bali. Landing during a prosperous harvest, they were hosted by the royal court. Seeing lush terraces and a wealthy king, the Europeans popularized the concept of a “Balinese Eden” [53].
They did not observe the system during a drought.
To build the lowland subak, people first dismantled the native environment. They cleared original forests and mixed wetlands, replacing them with an artificial landscape designed strictly for rice. Flooded paddies can temporarily support fish, frogs, snails, and aquatic insects, but the subak requires synchronized dry-downs. These dry periods eliminated the permanent water that native organisms needed to survive. The subak network redirected the entire local food web to serve human grain production, displacing native wildlife and severely shrinking the habitat of apex predators like the Bali tiger [54].
Once the natural wetlands were replaced by terraced paddies, water temples coordinated the irrigation.
These temples acted as a decentralized calendar, synchronizing exactly when massive blocks of farmland were flooded and when they were drained. This synchronization acted as a pest control system. When an entire district let its
fields dry out at the exact same time, the local rats and rice-eating insects starved. When the water returned, those populations had largely been killed, allowing farmers to grow massive amounts of rice without chemicals [55]. A single square kilometer of well-irrigated rice paddy can yield enough food for 300 to 700 people for a year [56]. That number only looks at the crop footprint. To function, the paddies required a significant supporting landscape. The Balinese left large upland forests untouched to catch the rain, prevent soil erosion, and supply the heavy amounts of firewood needed to boil the rice.
Bali also had a major geological advantage that kept productivity high for nearly a millennium. The island’s most productive sections drew their water from high volcanic crater lakes, like Mount Batur. The water flowing from these lakes carried a constant stream of dissolved volcanic phosphorus and potassium directly into the paddies [55]. By carving flat terraces into steep slopes and tapping into this gravity-fed, mineral-rich water, the Balinese farmed land more continuously than nature would normally allow. In its early centuries, Bali’s population hovered between 200,000 and 300,000 people. Factoring in the essential forests and watersheds, this resulted in an actual island-wide density of roughly 35 to 50 people per km². At this scale, the island produced enough food locally to sustain the population without foreign imports.
The society was at the mercy of the monsoon.
Because the subak’s greatest strength—synchronized
planting—required precise timing, any delay in the rains meant farmers collectively missed the planting window.
When droughts hit, the local food supply vanished. By the late sixteenth century, Dutch observers noted the king had to explicitly forbid rice exports during shortages. Famines routinely occurred [57].
Modern descriptions often characterize the subak as a grassroots democracy. However, in the fifteenth and sixteenth centuries, nobles and priests fleeing the collapse of the Majapahit empire in Java brought rigid political hierarchies to Bali. Irrigation became closely tied to dynastic power. Royal houses and noble lords controlled the massive labor forces needed to build and maintain the dams and canals [58, 59].
Warfare, elite extraction, debt, and slavery ran alongside Bali’s rice economy. Between 1620 and 1830, historians estimate that 100,000 to 150,000 Balinese were sold as enslaved people. In years of crisis, hungry and indebted peasants were seized and sold. The traditional story of the subak isolates the water system from the society that surrounded it, ignoring the political reality where rulers extracted grain, labor, and human lives [60].
As the terraces expanded and the population surged toward 900,000, surviving droughts required a massive stored surplus of grain. By this period the Balinese kings (rajas) engaged in regional maritime trade, exchanging agricultural surplus for foreign luxury goods like Indian textiles, Chinese ceramics, metal weapons, and opium. They
squeezed this surplus from the peasantry to fund imports and militaries, removing the emergency caloric buffer. The system was also physically vulnerable: a volcanic mudslide could instantly block an irrigation tunnel, and hostile armies routinely dammed rivers to starve out downstream rivals.
In 1815, the massive eruption of Mount Tambora on a nearby island pushed the region and subak system into total collapse. Volcanic ash wiped out the wet-rice crop [61].
With no stored surplus, the crop failures triggered severe famine, subsequent rat plagues, and epidemics. As the agricultural baseline collapsed, the elite accelerated the slave trade. Sir Stamford Raffles recorded that starving people and insolvent debtors were routinely enslaved, with male slaves selling for ten to thirty dollars [62]. Historian Anthony Reid noted that South Bali was so devastated it exported nothing but slaves, who were bought by French traders to work on sugar plantations in places like Réunion [63].
The rulers used the slave-trade profits to buy imported grain. But that food went to the royal courts and the military, while starving peasants were forced to clear miles of concrete-like volcanic ash from the terraces by hand.
In 1868, localized crop failures and a delayed monsoon killed roughly 2,000 people from starvation and disease [60]. By then, Bali and the broader Dutch Indies were importing hundreds of thousands of tons of grain annually from places like Burma and Siam [64]. By 1930, Bali’s population hit 1.1 million and was well past its years of self-sufficiency. Intense population pressure forced
farmers to subdivide their land into microscopic plots. By 1938, the average rice farm in Tabanan was just 0.61 hectares, leaving families reliant on external food sources [65].
In the 1970s, the Indonesian government rolled out a program called BIMAS (Mass Guidance) to increase yields.
Planners flooded the ancient terraces with synthetic nitrogen fertilizers, chemical insecticides, and high-yield rice varieties [55]. This industrialized farming approach replaced the ancient ecological loops of the subak.
Today, Bali holds over 4.3 million people—a density of more than 740 people per km². The island operates at a massive food deficit, kept alive by cargo ships and trucks importing tons of rice from Java and Lombok.
Tourism and housing development have pushed out of the valleys and up the volcanic slopes, destroying the mountain catchments. Now, wet seasons bring landslides, dry seasons bring water shortages, and roughly a thousand hectares of the ancient subak are paved over or abandoned every year [66].
The Nile Valley If you are looking for an environment capable of sustaining immense agricultural production for thousands of years, the Nile Valley is widely considered the ultimate success story.
“At present, it must be confessed, they obtain the fruits of the field with less trouble than any other people in the world… the husbandman waits till the river has of its own accord spread itself over the fields and withdrawn again to its bed, and then sows his plot of ground.” The Greek historian Herodotus made this observation around 440 BCE [67]. By the time of his visit, the inhabitants of the Nile Valley had already been farming the region for roughly 5,000 years. The nutrient-rich river flooded regularly, and thereby watered the fields and renewed the soil, doing work that Greek farmers often had to do themselves.
To build this agricultural center, the original ecosystem was destroyed. The society that took its place relied on physical labor, rigid social hierarchy, and an external footprint that extended well beyond the river valley.
Before human engineering, the Nile floodplain was a sprawling, biodiverse wetland. Ecological records show the valley was dominated by dense papyrus swamps and lotus marshes, serving as a thriving habitat for hippos, crocodiles, elephants, and millions of migratory birds [68].
To farm this landscape, ancient Egyptians spent centuries draining the wetlands, clearing native vegetation, and building a vast network of earthen levees to plant a monoculture of emmer wheat and barley. As the farms expanded, the native ecosystem was eradicated, driving the
local megafauna extinct or pushing them far south [69].
This ecological shift had two severe consequences. First, by eliminating the diverse wild food buffer (fish, fowl, and marsh plants), the population became entirely dependent on the exact height of the annual river flood. When floods were weak for multiple years, famine followed.
Second, replacing flowing wetlands with stagnant, shallow irrigation basins created a permanent public health crisis. Modern bioarchaeology shows that the shift to basin irrigation triggered endemic rates of schistosomiasis—a waterborne parasitic flatworm disease that causes chronic fatigue, organ damage, and early death. Traces of the parasite are found in Egyptian mummies across all social classes [70].
The agricultural system also required immense physical labor. Maintaining thousands of miles of canals and dikes required an unpaid, forced labor draft. Every year, during the flood season, the state forced peasant farmers (the fellahin) to dredge thick mud, rebuild earthen walls, and haul water entirely by hand under the threat and use of physical violence.
Skeletal remains from workers’ cemeteries reveal the toll of this labor. Bioarchaeologists note high rates of osteoarthritis and spinal degeneration from heavy load-bearing. Coupled with high infant mortality, the average life expectancy for the working class hovered around 30 years [71].
Despite the human cost, the system produced a staggering amount of food. It survived for millennia
without modern chemical fertilizers thanks to a continent-sized geological subsidy: the Ethiopian monsoon.
Every summer, torrential rains battered the Ethiopian Highlands, carving millions of tons of mineral-rich volcanic topsoil out of the mountains. The Blue Nile carried this silt thousands of miles north, depositing it perfectly across the Egyptian floodplain during the annual flood [72]. This fresh sediment allowed farmers to repeatedly grow crops in the same soil, supplemented by crop rotation, legumes, and fallowing [73].
The sheer volume of food produced in the Nile Valley often leads to a misunderstanding about how ancient Egypt functioned. At its peak in the New Kingdom (roughly 1550–1070 BCE), the core floodplain supported about 3 million people [71]. Because the farmable land covered roughly 24,000 square kilometers, it is frequently assumed the society efficiently sustained a dense 120 people per km².
This high residential density is not proof of self-sufficiency.
Ancient Egypt was never a closed system. While the valley provided enormous amounts of grain, it lacked the other resources required to maintain a complex civilization.
Farming at a state level requires timber, tools, cargo ships, transport corridors, copper, and more. The lack of domestic timber was a matter of basic survival. To prevent localized droughts from causing mass starvation, the state had to
transport thousands of tons of surplus grain up and down the river. They could not build 50-ton cargo barges out of native Egyptian trees, which yield short, brittle wood. To survive, the state had to import thousands of tons of cedar from the mountains of Lebanon [74, 75] and mine copper from the Sinai.
When this external supply chain broke—such as during the collapse of the First Intermediate Period—the transport network failed, and the centralized granaries could not distribute food. Primary sources from the era, such as the tomb inscription of the provincial governor Ankhtifi, document mass starvation, noting that “all of Upper Egypt was dying of hunger, to such a degree that everyone had come to eating his children” [76]. The 24,000 km² Nile Valley was never enough; the “ghost acreage” of external forests and mines was structurally required to keep the system from starving itself.
To permanently secure these resources, the later New Kingdom conquered them. While the political map expanded across roughly 500,000 km² of the Levant, Sinai, and Nubia, much of that was uninhabitable desert where no extraction occurred. The utilized ecological region—the core floodplain, the arable coastal strips and forests of the Levant, the narrow Nubian river corridor, and the active mining zones—amounted to roughly 65,000 to 80,000 km².
When you divide the total population of the empire (roughly 4.5 to 5 million) by the land actually required to
sustain it, the population density of the society drops to roughly 55 to 75 people per km² [77].
Once Rome conquered Egypt in 30 BCE, the region ceased to be an independent empire and was absorbed into a Mediterranean supply chain. Because Rome managed the external resources (like timber and metals), Egypt was forced to hyper-specialize as an imperial granary.
The population of the core floodplain swelled to nearly 5 million people [78].
To feed its capital, Rome demanded an enormous annual tax extracted directly in wheat—roughly 135,000 tonnes a year [79]. This extraction stripped away the local population’s emergency caloric buffer. When the Nile flood was low, Rome did not lower its quotas; the grain ships still sailed, and the local farmers were left with nothing.
Administrative records document a recurring crisis known as anachoresis (flight). Starving farmers abandoned their villages and fled into the desert to become bandits rather than face torture for failing to meet the imperial grain quotas [80].
Ultimately, this peak density proved unsustainable.
The intricate irrigation system relied on highly coordinated labor. When the Roman and Byzantine administrations fractured, and plagues decimated the workforce, the canals choked with silt [81]. Without maintenance, the river’s geological subsidy couldn’t reach the fields, and the population crashed. By the time Napoleon surveyed the
country in 1798, Egypt’s population had fallen back to roughly 2.5 to 3 million [82].
In the twentieth century, the ancient agricultural rhythm was permanently severed. To feed a rapidly growing population and to grow cash crops like cotton year-round, the government abandoned seasonal basin irrigation. This shift culminated in the 1970 completion of the Aswan High Dam [83]. The dam granted total control over the river, but it ended the 5,000-year fertilizer supply. The annual flood of volcanic silt is now trapped behind concrete in Lake Nasser.
Deprived of those nutrients, the soil of the Nile Delta began to rapidly degrade and salinize [84].
Today, the same floodplain holds over 110 million people. They no longer survive on the natural carrying capacity of the Nile. Modern Egypt is now one of the highest consumers of synthetic nitrogen fertilizers on Earth [85] and operates as one of the world’s largest importers of foreign wheat [86].
Edo Period Japan Edo-period Japan (1603–1868) is frequently invoked by environmentalists as a documented blueprint for survival.
Designer Azby Brown points to the era as a “model of how to flip impending environmental collapse into sustainability,” noting that limited resources forced an integration of environmental consciousness into everyday design and agriculture [87]. It is celebrated as a historical triumph of the circular economy; researcher Eisuke Ishikawa
describes it as “a nearly perfect circulation type society” where “virtually everything was circulated in loops” through the rigorous reuse of ash, textiles, night soil, and everyday materials [88]. For those seeking to answer if a large agricultural population can survive indefinitely without fossil fuels, Edo Japan is often presented as the premier example. Advocates such as chemist Ei-Ichiro Ochiai claim the system operated “virtually without input of energy and material from the outside; i.e., depending solely on solar energy” [89].
The era began with a march toward strict geographic isolation. After unifying the country in 1603, the Tokugawa shogunate spent three decades restricting foreign contact. Under the Sakoku (closed country) edicts, foreign entry was strictly limited, and overseas trade was monopolized by the state and restricted to a handful of designated gateways. To survive, the nation had to rely almost entirely on what its own islands and adjacent waters could produce [90].
When this isolation began, the Japanese population was roughly 12 to 15 million people. Spread across the archipelago, this yielded a density of roughly 30 to 40 people per km² [91]. At this scale, the agricultural system might have been sustainable, but the population did not remain there long enough to find out.
To feed themselves, the Japanese relied on wet-rice agriculture. Because Japan is overwhelmingly mountainous, only about 15 percent of the landmass—roughly 60,000
square kilometers—was flat enough to farm. Growing rice without fossil-fuel fertilizers requires significant nutrient inputs. The Japanese famously utilized a “night soil” loop, returning urban human waste to the fields as nitrogen.
While often praised as a perfect circular economy, cycling human feces directly into the food supply carried a biological cost. The practice caused endemic rates of soil-transmitted parasites, specifically roundworm and whipworm, which infected the population and consumed a portion of their already limited caloric intake. Furthermore, human waste alone cannot replace the total organic matter stripped away by a heavy harvest. Keeping the soil alive required a massive secondary footprint. Historical agronomy shows that to maintain just one hectare of intensive rice paddy, a village needed the satoyama, five to ten hectares of surrounding wild hillside. Every year, peasants had to hike into the mountains, harvest grasses, leaves, and young saplings, and carry the loads down to the valleys to use as kariyashiki or green manure [92].
This continuous extraction degraded the mountain soils and kept the upland forests in an arrested state.
Stripped of their deep-rooted biological sponge, the mountainsides became highly unstable. During typhoons, the hills collapsed, unleashing landslides and lowland flooding that buried the rice paddies. To prevent total deforestation and stabilize the hydrology, local lords seized control of the woodlands. They severely restricted peasant
access to the mountains and enforced logging bans with capital punishment [92].
Locked out of the mountains, the Japanese turned to the sea. They dredged the ocean for fertilizer, harvesting millions of tons of dried sardines. To keep the soils of the main islands alive, the shogunate extended its biological footprint to the north, extracting large amounts of herring from Indigenous Ainu lands in Ezochi (modern Hokkaido) to boil into fishmeal [93].
Under this expanding extraction system, the population doubled, reaching roughly 30 million by the 1720s. At 30 million, the islands hit a biological wall. For the next 150 years, the population stagnated [91].
If you look only at the 60,000 square kilometers of flat farmland, supporting 30 million people creates the illusion of a density of 500 people per km². Keeping that farmland alive required the five-to-one ratio of mountain catchment. That required roughly 300,000 square kilometers of mountains—slightly more land than existed on Japan’s three main islands. The Tokugawa system maxed out the physical geography of the main islands and spilled over into the sea and Hokkaido to make up the biological deficit. When you measure the peak population against the entire utilized archipelago, the absolute biological ceiling of the system was roughly 80 people per km² [94].
This was far from sustainable. And temporarily maintaining society at this biological maximum required rigid social control. Unlike rainfall-dependent crops,
wet-rice agriculture relies on a shared, fragile hydraulic network. Every terraced paddy is physically connected by a continuous flow of water running through an intricate system of hand-dug canals and earthen embankments. As agricultural historian Francesca Bray documents, independent farming under these conditions is physically impossible; “a single breach in a dike can ruin a whole hillside of terraces,” dictating “strict social regulation” and synchronized community labor to manage the water grid [95].
The Tokugawa shogunate enforced this coordination by binding the peasantry into goningumi—five-household groups linked by mutual surveillance and collective responsibility [96]. If one peasant fled to the city, abandoned their fields, or defied a decree, the other four households were punished. To prevent social mobility, the government micromanaged peasant lives down to their daily calories. The famous Keian Edict of 1649 explicitly forbade peasants from eating the rice they grew, ordering them to subsist on “millet, vegetables, and other coarse food,” while prohibiting them from buying “sake or tea” [97].
When the weather shifted or volcanoes erupted, the tight ecological margins failed. The resulting famines—most notably the Kyōhō (1732), Tenmei (1780s), and Tenpō (1830s)—killed millions through starvation and resulting disease. To keep from starving during normal years, peasants were forced to manage their own numbers through mabiki.
Translated literally as “thinning out the seedlings,” mabiki was the widespread practice of infanticide. Families knew exactly how many mouths their land could support, and midwives routinely ended the lives of newborn infants to keep family sizes within the strict limits of the exhausted soil [88].
The administrative doctrine of the era was ikasazu korosazu, to assess taxes so heavily that the peasants “can neither live nor die.” Another popular administrative maxim of the period was equally blunt: “Peasants are like sesame seeds: the more you squeeze them, the more you get out” [98].
The samurai class functioned as heavily armed overseers with the legal right of kiri-sute gomen—the right to strike and abandon. Article 71 of the Legacy of Tokugawa Ieyasu, which codified the shogunate’s rules, explicitly authorized lethal force against commoners for breaches of etiquette: “The samurai are the masters of the four classes.
Agriculturists, artisans and merchants may not behave in a rude manner towards samurai… and a samurai is not to be interfered with in cutting down a fellow who has behaved to him in a manner other than is expected.” [99] When a daimyō traveled the roads, commoners were required to perform dogeza—dropping to their knees and pressing their faces directly into the dirt. German physician
Engelbert Kaempfer, traveling through Japan in the 1690s, documented these processions, noting that anyone on the road “must prostrate themselves flat on the ground… and not presume to look up” [100].
By the mid-19th century, the system was failing.
The forests were depleted, the peasantry was staging uprisings, and the isolationist government was crumbling.
When American Commodore Matthew Perry arrived with a fleet of steamships in 1853 demanding trade, the fragile social order shattered.
The physical subjugation required to maintain the agrarian system sparked a war when it was applied to the outside world. On September 14, 1862, a British merchant named Charles Lennox Richardson was riding his horse near Edo when he encountered the procession of the Satsuma regent. Unused to the absolute deference demanded of the Japanese peasantry, Richardson failed to dismount and prostrate himself. Treating this as routine agrarian crowd control, the samurai guards drew their swords and killed him [101]. When the British demanded reparations and the Japanese refused, the Royal Navy bombarded the city of Kagoshima.
Following the Meiji Restoration in 1868, Japan rapidly modernized and transformed into an imperial power. To secure the fertilizer they desperately needed, Japan formally annexed Hokkaido, accelerating the extraction of its marine wealth, and soon expanded into Taiwan and Korea, conquering new lands to secure the
sugar, rice, and physical space their own islands could no longer provide [102].
Today, Japan crowds more than 124 million people onto those same islands. That is a population density of 338 people per km² [103]. With the modern government warning that society will collapse if citizens do not have more children, the historical amnesia is total. As Prime Minister Fumio Kishida declared in 2023, the declining birth rate means “Japan is standing on the verge of whether we can continue to function as a society” [104].
The Iberian Dehesa The Iberian dehesa (known as the montado in Portugal) is frequently cited by agricultural scientists and the European Union as a model of humans living in balance with nature.
Spanning over 3 million hectares across southwestern Spain and southern Portugal, it is a multi-functional oak savanna.
Ecologically, the dehesa prevents soil erosion in an arid climate, requires minimal chemical fertilizers, and preserves ancient oaks that would have been removed by deep-plow agriculture. Because it supports biodiversity while functioning as an agricultural system, the EU officially designates it as a “High Nature Value” (HNV) farming system [105].
The dehesa was one of the most efficient, low-input, high-yield export engines in pre-industrial Europe. It was not designed as a subsistence landscape, but as a producer of luxury cash crops for global markets. Its
original function was serving as winter pasture for Spain’s transhumant Merino sheep. During the 15th and 16th centuries, the Concejo de la Mesta (the national guild of sheep owners) oversaw a flock of roughly 2.5 to 3 million Merino sheep that migrated annually to the southern dehesas [106]. This produced the highest-grade wool in Europe. At its peak in the mid-16th century, Castile exported roughly 4 to 5 million pounds of raw Merino wool annually to the textile hubs of Flanders, England, and Italy [107].
This wool trade was the backbone of the Spanish imperial economy. The Crown collected massive revenues through the servicio y montazgo (transit and grazing taxes).
The dehesa landscape was so lucrative that the Crown granted the Mesta legal privileges to protect it, strictly banning the conversion of pasture into plowed farmland [106]. As the wool trade declined in the 19th century, the dehesa transitioned to other luxury exports, such as cork and high-end pork.
The dehesa is not a natural landscape; it is a highly engineered ecosystem. Before human intervention, the Iberian southwest was dominated by dense climax forests of holm oak, cork oak, and thick scrub. Beginning in the Middle Ages, landowners thinned the canopy and eradicated the undergrowth to create open pastures for the Merino flocks. Ecologists define the resulting landscape as an “arrested successional stage” [108].
These grasslands allowed open-woodland and edge-habitat species to thrive. However, the clearing displaced the original deep-forest species, including ambush predators, brush-nesting birds, and moisture-dependent amphibians. Without ongoing grazing and clearing, the dehesa rapidly reverts to dense scrubland, and the open-woodland biodiversity vanishes [108].
Because the dehesa was optimized for export, its capacity to produce local food for human consumption was restricted. From the Middle Ages through the 18th century, the regions defined by the dehesa recorded low population densities of roughly 10 people per km². This was a deliberate feature of the model. To maximize unbroken pasture for millions of sheep, the workforce was not permitted to live in dispersed farmsteads across the territory. Instead, the peasantry was concentrated into massive, isolated nucleated settlements. These towns were surrounded by enormous latifundios (private estates) where jornaleros (landless laborers) were permitted to work seasonally but legally barred from living on or cultivating the land for their own subsistence [109].
Because the land was locked up as pasture, the region was structurally food-insecure. While the estates generated wealth through wool, cork, and meat, the land did not yield baseline calories for the local laborers. Local workers survived on grain imported from northern Castile or via sea trade, making staple foods artificially expensive.
The historical diet of the jornalero consisted almost entirely
of bread, olive oil, garlic, and vinegar—the origin of gazpacho as a basic caloric ration. When droughts hit the northern wheat belts and grain prices spiked, the dehesa regions experienced severe “price famines,” with major crises recorded in 1868, 1882, and 1904 [109].
Maintaining this landscape without modern machinery required an army of manual labor. Spanish Enlightenment reformer Gaspar Melchor de Jovellanos documented this in his 1795 Informe sobre la Ley Agraria, noting that the legal protection of pastures artificially restricted the food supply while the rural workforce lived in systemic poverty [110]. The landscape’s balance was maintained largely because human labor was inexpensive, and laborers had no legal right to use the land to feed themselves.
Today, the region produces roughly 80 percent of the world’s commercial cork, harvested from the bark of the Quercus suber without killing the trees [111]. The acorn drop from the holm oaks fattens the Iberian pig, producing premium luxury exports. However, the socioeconomic structure that maintained this landscape for centuries no longer exists.
Landowners have turned to heavy mechanization—which compacts the soil and damages tree roots—and have raised livestock densities, leading to overgrazing. Ecologists warn of a critical “lack of regeneration” in the dehesa [112].
Simultaneously, the aging canopy is dying off from la seca, a devastating root rot caused by the soil-borne pathogen Phytophthora cinnamomi, which thrives in compacted soils and preys on stressed, mechanically damaged trees [113].
The Iberian dehesa is a historically powerful engine for generating luxury export wealth, but it is not a model for how a dense human population can sustainably feed itself. It remains an artifact of an export economy that prioritized high-value commodities over local carrying capacity, maintained for centuries by the seasonal labor of a landless underclass.
Reaching for Maximum Sustainability The floating gardens of the Aztecs, the water temples of Bali, the floodplains of the Nile, the closed-loop fields of Edo Japan, and the engineered savannas of the dehesa demonstrate that centuries of agricultural continuity are possible. And they reveal a stark historical reality. These hyper-productive landscapes also produced famine, slavery, and ecological destruction. They exacted a heavy human toll through rigid labor hierarchies, relied upon rare geological blessings, and depended on the wealth of sprawling empires to manage the thin boundary between perpetual renewal and systemic collapse.
We have now reached the hardest part of the chapter.
Our first estimate gave us a baseline: roughly 40 to 170 million people. That number came from mapping habitable land, broad habitat categories, and the historically observed densities of people living from ecosystems that kept renewing themselves.
Carrying capacity papers often ignore sustainability, then look at the highly engineered landscapes of the Aztec chinampas, the Balinese subak, or Edo Japan, and build a model that scales their massive densities globally.
If we follow that logic, we would assume that 104 million km² of habitable land could be converted to these types of agricultural systems. Using population densities that appeared close to sustainable for those systems, ~30 to 50 people per km², we could calculate a “sustainable” global population in the 3–5 billion range.
This would be wildly inaccurate, as we have just walked through in detail.
For example, the Nile Valley is a geographic rarity, but there are similar regions that have been gifted with nutrient-rich rivers that regularly flooded the nearby land.
Noncoincidentally, these also were cradles of early civilization. The Tigris and Euphrates rivers (Mesopotamia), the Yellow River (North China Plain), and the Indus river (The Punjab, The Sindh). The densely populated Mesopotamia and North China Plain largely resemble the substantial forced labor and famine history of Egypt.
The Indus valley civilizations appear to be the exception. We will return to that shortly.
The combined peak populations these regions reached is 70–75 million people, and at these levels the societies were highly unsustainable. It is completely inappropriate to export these already buckling densities from the most fertile regions on the planet and apply them in tundra, semi-arid land without a regularly flooding, nutrient-rich river, boreal forest, prairies, savannas, or dense forests.
It also ignores all of the slavery.
Some may argue that slavery is sustainable and should be used to estimate human carrying capacity. I will leave them to do that analysis.
As an advocate of freedom, I instead recommend we look at what large-scale agriculture looks like when slavery appears to be absent.
The Indus Valley Civilization (IVC) spanned roughly 800,000 square kilometers across modern Pakistan and northwestern India and built dozens of planned cities.
However, the archaeological record of the IVC lacks the physical evidence of coercive states found in Egypt, China, or Mesopotamia. Excavations have not uncovered monumental royal tombs, large armories, or artwork depicting bound captives and slaughtered enemies [114].
Instead of central palaces, the society’s infrastructure focused on public utilities. Most of the urban population lived in standardized, multi-story baked-brick
homes with private wells and access to an enclosed, city-wide sewer system [115]. Measuring wealth disparities through house sizes, archaeologist Adam Green notes that the IVC operated with “trivial inequality” compared to contemporaneous cities. Historian Gregory Possehl described the civilization as achieving “sociocultural complexity without the state.” Governance appears to have relied on civic coordination, corporate trade guilds, and economic regulation rather than military force [116].
Based on the geographic spread of the civilization, mid-century estimates placed the population at up to 5 million people. More recent spatial analyses of the physical settlements have revised that number downward. Farmers in the Indus Valley lived clustered in settlements rather than in isolated structures spread across the plains. Archaeologists have mapped roughly 1,900 small farming villages (averaging 2.5 hectares), roughly 100 provincial towns, and five primary urban centers [117]. When mapping the agricultural base directly to these physical hectares and applying standard pre-industrial settlement densities, the total population of the IVC is estimated at 1.2 to 1.5 million people [118].
Spread across their 800,000 km² footprint, this yields an average density of 1.8 people per km². They sustained this peak for roughly 700 years and farmed the region for 2,000 years.
This density is significantly lower than the agricultural concentrations of the Nile Valley and falls below
the density of the most highly productive coastal forager environments. The society maintained its population by spreading out across a highly networked landscape.
This still required substantial environmental engineering. The baseline carrying capacity of the semi-arid environment—a blended average of the river corridors and dry scrublands—could naturally support roughly 0.3 foragers per km². By using irrigation to expand the farming corridors and domesticated cattle to extract calories from the dry savanna, the civilization supported roughly six times the environment’s natural baseline.
We are still unable to decipher their writing so we cannot say for sure, but the evidence suggests they had neither slaves nor masters.
How did they avoid the despotic rulers and coercive labor systems that dominated their contemporaries?
Anthropologists point to two physical realities of the Indus Valley: unconstrained geography and decentralized resources. In regions like Egypt, farmers were geographically trapped between the Nile and impassable deserts; if they refused the state’s demands, they had nowhere to flee. The Indus Valley, by contrast, was a sprawling, highly networked environment. If an aspiring elite attempted to extract forced labor, the population had an exit strategy—they could simply migrate along the river systems and establish new settlements [119].
Furthermore, the region contained abundant, locally accessible copper. In states like Mesopotamia, elites
maintained power by controlling the rare supply of imported metals, which allowed them to monopolize military force. In the IVC, metalworking was highly decentralized, taking place in small workshops across residential neighborhoods. Because metal tools were widely distributed and the population was not geographically trapped, the physical mechanics of subjugation were highly impractical. Without the ability to monopolize force, the civilization relied on economic integration rather than the threat of a royal army [120].
The Indus Valley demonstrates that a complex, productive society can organize without evidence of widespread forced labor. However, the society remained bound by environmental limits.
Maintaining the cities required extensive resource extraction. Firing the millions of baked bricks used in construction required significant amounts of timber, which paleoenvironmental models suggest led to localized deforestation and the degradation of the surrounding riparian wetlands [121].
Furthermore, the agricultural surplus that sustained the population was dependent on a specific Holocene climate pattern of highly predictable summer monsoons.
Around 1900 BCE, ocean temperatures shifted, and the monsoons weakened [122].
As the agricultural base failed, the population exceeded the changing carrying capacity of the environment.
Skeletons from the decline phase show a marked increase in
lesions caused by severe childhood anemia and chronic malnutrition. As nutritional stress increased, the skeletal record shows a concurrent rise in infectious diseases.
Bioarchaeologist Gwen Robbins Schug notes that during this period of climate uncertainty, “the prevalence of infection and infectious disease increased through time,” including early strains of tuberculosis and leprosy [123].
The archaeological record indicates that the civic cohesion of the preceding centuries broke down. According to Robbins Schug, the Late Harappan period shows a sharp increase in cranial trauma, supporting “a growing pathology of power” and interpersonal violence as the resource base contracted [124]. Without the agricultural surplus to sustain them, the civilization de-urbanized, and the cities were abandoned.
A society does not need an emperor or an enslaved underclass to hit its carrying capacity. It only needs to exceed its natural subsidy.
Pushing the global population to its limits has clear downsides. The Indus Valley was able to sustain a large population and then it wasn’t. A society maximizing for the current moment is a population at higher risk for famine and collapse than one that aims for a buffer. When temperatures, rivers, resources, pest resistance, politics, rain cycles, monsoons, or weaponry shift, then the carrying capacity shifts.
The IVC were able to reach around 6 times what their region would have supported without their
innovations and significantly degraded their environment.
For most of their existence, their skeletons show a far healthier existence than their enslaved peers in other regions.
The Indus Valley suggests that my baseline range of 40–170 million does not fully account for human ingenuity.
If we overcorrect and apply the Indus Valley’s ingenuity multiplier to our highest baseline densities across all habitat zones, we reach a high-end estimate of a 1.02 billion person population. This is not an estimate for a carrying capacity, but a population density sanity check.
We can use 40 million, 170 million, 1 billion, or 2 billion and the answer remains the same: we have billions more people today than we can sustainably support.
Today, the Earth holds 8.3 billion people on roughly 104 million square kilometers of habitable land.
That is a global average of roughly 80 people per km². We are actively attempting to live at the precarious peak density of Edo Japan, but on a planetary scale. The resources enabling this overshoot are disappearing faster than we can replace them. The fossil fuels, the primary driver of all of this overshoot, are not being replaced at all. We cannot sustain another century of this. A living Earth is not a static machine, and humans are capable of incredible ingenuity.
We can participate in ecosystems. We can tend forests, manage fire, build soil, and restore watersheds. A society that successfully lowers its pressure on the Earth can dedicate its energy to true innovation. It can learn how to increase the planet’s carrying capacity by making the living
world richer, rather than just extracting from it. Future generations may discover how to build highly productive, sustainable food systems that do not require either ecological destruction or human slavery.
That is work to be done. It is not a number we can assume today to justify four, eight, or ten billion people.
The goal of human civilization should not be to calculate the absolute maximum number of breathing bodies we can cram onto a dying planet. I prefer to solve for freedom, biological abundance, and sustainability.
We are nowhere near that balance.
To get there, we need fewer people.