The Last Drop: Why Every Civilisation That Ran Out of Water Collapsed

In January 2018, the city of Cape Town announced that it was ninety days away from running out of water. Not running low. Not facing shortages. Running out entirely. The municipal government published a countdown — “Day Zero” — the date on which the taps would be turned off and four million residents would be directed to collect a daily ration of twenty-five litres from two hundred military-guarded distribution points across the city. Twenty-five litres. That is roughly one flush of a Western toilet and a brief shower. For everything — drinking, cooking, washing, sanitation — for an entire day.

Day Zero, ultimately, did not arrive. A combination of severe water restrictions (residents were limited to fifty litres per day, down from the South African average of two hundred and thirty-five), emergency groundwater drilling, and the belated arrival of winter rains pulled Cape Town back from the brink. But the crisis laid bare a truth that most modern cities prefer not to think about: the water coming out of your tap is not guaranteed. It never has been. And for a growing number of cities around the world — Chennai, São Paulo, Mexico City, Jakarta, Cairo — it is becoming less guaranteed with each passing year.

The Romans built aqueducts that still work after two thousand years. We built cities in deserts and act surprised when the taps run dry.

This is a History Future Now article. So we are going to look at the present through the lens of the deep past, and then project forward. The question is straightforward: what is the relationship between water and civilisation? The historical record provides an answer that is as clear as it is uncomfortable. Every major civilisation in history was built on the mastery of water. And every civilisation that lost control of its water supply collapsed. Not declined. Not adapted. Collapsed.

The modern world, with its dams, desalination plants, and deep-bore pumps, has convinced itself that it has transcended this ancient dependency. The evidence suggests otherwise.

The Hydraulic Civilisations

In 1957, the German-American historian Karl Wittfogel published Oriental Despotism, a sprawling and controversial work that proposed what became known as the “hydraulic hypothesis.” Wittfogel’s argument was deceptively simple: large-scale irrigation requires centralised coordination; centralised coordination requires bureaucracy; bureaucracy requires taxation; taxation requires a state. Water management, in other words, did not merely sustain early civilisations — it created them. The state itself was, in Wittfogel’s telling, a by-product of the need to dig canals.

The thesis has been refined and challenged in the decades since, but the core observation remains remarkably robust. The earliest complex societies on Earth — Mesopotamia, Egypt, the Indus Valley, China — all arose in arid or semi-arid regions where agriculture was impossible without the collective management of water. This is not a coincidence. It is a pattern so consistent that it amounts to something close to a historical law.

Mesopotamia — the land “between the rivers” — is the canonical example. The Tigris and Euphrates provided water, but not reliably. The rivers flooded unpredictably, and the flat alluvial plain between them required an elaborate network of canals, levees, and reservoirs to distribute water to fields. The Sumerians, and after them the Akkadians and Babylonians, built these systems at a scale that boggles the modern mind. By the third millennium BC, the city of Girsu alone maintained a canal network irrigating over 4,000 hectares (Jacobsen and Adams, 1958). Managing this infrastructure required record-keeping (hence cuneiform, one of the world’s first writing systems), centralised administration, corvée labour, and the taxation to fund it all. The canal did not serve the state. The canal was the state.

Egypt took a different approach. The Nile’s annual flood was remarkably predictable — it rose in June, peaked in September, and receded by November, depositing a layer of rich silt across the floodplain. But predictable is not the same as manageable. The Egyptians developed the Nilometer — a series of stone gauges built into the riverbank at key points, including on Elephantine Island near modern Aswan — to measure the flood’s height. Too low a flood meant famine. Too high meant destruction. The Nilometer readings determined tax assessments for the year: a good flood meant a good harvest, which meant higher taxes could be collected (Said, 1993). The pharaoh’s legitimacy rested, in no small measure, on his perceived ability to mediate between his people and the river. When the floods failed, dynasties fell.

The Indus Valley civilisation, centred on the great cities of Mohenjo-daro and Harappa (c. 2600–1900 BC), provides perhaps the most sophisticated example of early hydraulic engineering. Mohenjo-daro’s Great Bath — a watertight pool measuring twelve metres by seven, lined with bitumen — was almost certainly a ritual or public bathing facility, but the city’s true engineering marvel was its drainage system. Every house had a bathroom connected to a covered street drain, which flowed into a main sewer. This was sanitary infrastructure that European cities would not match for another four thousand years. The Indus cities also maintained an elaborate system of wells, reservoirs, and water channels that served a population estimated at forty thousand or more per city (Kenoyer, 1998).

China completed the Grand Canal — at nearly 1,800 kilometres, the longest artificial waterway in history — in sections over more than two thousand years, beginning in the fifth century BC and reaching its full extent under the Sui Dynasty in the seventh century AD. The canal linked the Yangtze River in the south to the Yellow River in the north, enabling the transport of grain from the fertile rice-growing regions to the arid political capitals. It was, in effect, a man-made river that held together an empire. When sections silted up or fell into disrepair, famine and rebellion followed with mechanical regularity (Needham, 1971).

The pattern across all four cases is identical. Water management required collective action on a scale that no family, village, or tribe could achieve alone. This collective action demanded coordination, which demanded hierarchy, which demanded records, which demanded literacy, which demanded a professional administrative class — which is to say, a government. Wittfogel overstated the determinism, and his political conclusions about “Oriental despotism” have not aged well. But his central insight — that water was the organising principle of early civilisation — remains one of the most powerful ideas in historical sociology.

When the Water Stopped

If water management built civilisations, then the loss of water destroyed them. The archaeological record is littered with the remains of societies that thrived for centuries and then, within a few generations, vanished. In a striking number of cases, the collapse correlates with a disruption to the water supply — whether from drought, over-irrigation, salinisation, or the failure of hydraulic infrastructure.

The Akkadian Empire provides the earliest well-documented case. Founded by Sargon of Akkad around 2334 BC, it was arguably the world’s first empire, unifying the Sumerian city-states under a single authority that stretched from the Persian Gulf to the Mediterranean. It lasted barely a century. Around 2200 BC, the empire disintegrated with a speed that has puzzled historians for generations. Archaeological evidence from Tell Leilan in modern Syria shows that the city was abruptly abandoned — not sacked, not burned, but simply left empty, with a thick layer of wind-blown dust accumulating over the ruins (Weiss et al., 1993).

The explanation came from climate science. Peter deMenocal, analysing deep-sea sediment cores from the Gulf of Oman, demonstrated that the period around 2200 BC coincided with a sharp increase in wind-borne dust — a proxy for severe arid conditions across Mesopotamia. The drought was not a brief dry spell. It lasted approximately three hundred years (deMenocal, 2001). The irrigation canals that had sustained the Akkadian agricultural system could not function without river water, and the rivers depended on rainfall in the headwaters. When the rain stopped, the canals dried up, the harvests failed, and the administrative apparatus that held the empire together lost its reason for existing. The population scattered. The cities emptied. The first empire in history was undone by a change in rainfall patterns.

The Classic Maya collapse of the ninth century AD follows an eerily similar script. At its height, the Maya civilisation supported a population of perhaps ten to fifteen million across the Yucatán Peninsula, Guatemala, and Honduras. The cities — Tikal, Calakmul, Copán, Palenque — were architectural marvels, with elaborate systems of reservoirs, canals, and terraced hillsides designed to capture and store seasonal rainfall in a region with no permanent rivers and a pronounced dry season. Then, between roughly 800 and 1000 AD, the major southern lowland cities were abandoned. Population collapsed by perhaps seventy to ninety per cent (Webster, 2002).

Hodell, Curtis, and Brenner, analysing sediment cores from Lake Chichancanab in the Yucatán, found evidence of the most severe drought in the region in seven thousand years, peaking between 800 and 1000 AD (Hodell et al., 1995). Subsequent research has confirmed and refined the picture: the Maya collapse coincided with a series of prolonged droughts, each lasting years to decades, punctuated by brief wet intervals that offered temporary respite before the next dry period (Kennett et al., 2012). The Maya had built their civilisation on the assumption that the rains would continue. When they did not, the reservoirs emptied, the crops failed, and the political structures that depended on surplus food production disintegrated.

The Khmer Empire of Southeast Asia offers a variation on the theme. Angkor, the Khmer capital, was the largest pre-industrial city in the world, covering an area of roughly one thousand square kilometres — larger than modern-day Los Angeles. Its population at its peak in the thirteenth century may have reached seven hundred and fifty thousand or more (Evans et al., 2007). The city was sustained by an extraordinary water management system: an intricate network of canals, dykes, and massive reservoirs (called barays) that captured monsoon rainfall and distributed it across the rice paddies throughout the dry season. The West Baray alone held more than fifty million cubic metres of water.

But the system was fragile. Archaeological and hydrological analysis using LIDAR mapping and sediment cores has shown that by the fourteenth and fifteenth centuries, the canal network was silting up, sections were being deliberately breached (possibly during factional conflicts), and the entire hydraulic system was deteriorating under the strain of a growing population and shifting monsoon patterns (Buckley et al., 2010). When the Siamese kingdom of Ayutthaya sacked Angkor in 1431, the city was already in terminal decline. The invaders delivered the final blow, but the water system had failed first.

The American Dust Bowl of the 1930s is the modern world’s nearest equivalent. Between 1930 and 1940, a combination of severe drought and catastrophic land management turned the Great Plains of the United States into a wasteland. Topsoil, exposed by decades of aggressive ploughing that had destroyed the native grassland, was carried away in apocalyptic dust storms — some reaching as high as three thousand metres and travelling as far as Washington, DC. An estimated 2.5 million people fled the affected states (Worster, 1979). It was the largest peacetime migration in American history, a humanitarian disaster in the richest country on Earth, caused by the oldest mistake in the book: assuming the water would always come.

The pattern is so consistent that it ceases to be coincidental. Civilisations that master water thrive. Civilisations that lose control of water — through climate change, over-extraction, infrastructure failure, or sheer negligence — collapse. There are no exceptions.

The Modern Crisis — Who Is Running Dry?

The comforting assumption of the modern era is that technology has liberated us from the ancient dependency on rainfall and rivers. We have dams, pumps, pipes, treatment plants, and desalination facilities. We can move water across continents. We can pull it from aquifers a kilometre beneath the surface. Surely the lesson of the Akkadians and the Maya is interesting but no longer relevant?

The data says otherwise.

The Colorado River, which supplies water to forty million people across seven US states and northern Mexico, no longer reaches the sea. It has not done so reliably since the 1960s. Lake Mead, the largest reservoir in the United States, fell to 27 per cent capacity in 2022 — its lowest level since it was filled in the 1930s. Lake Powell, the second-largest, dropped to 24 per cent. The Bureau of Reclamation declared the first-ever federal water shortage on the Colorado in August 2021, triggering mandatory cuts to Arizona, Nevada, and Mexico (Bureau of Reclamation, 2021). Seven states, two countries, forty million people, and a river that is being consumed faster than the Rocky Mountain snowpack can replenish it.

The Aral Sea, once the fourth-largest lake in the world, has lost approximately ninety per cent of its volume since 1960. Soviet-era diversion of the Amu Darya and Syr Darya rivers for cotton irrigation turned a vast inland sea into a toxic dustbowl. The eastern basin dried up completely in 2014. The fishing industry that once employed sixty thousand people is gone. The exposed lakebed — contaminated with pesticides and salt — generates dust storms that spread respiratory disease across the region (Micklin, 2007). It is the single largest man-made environmental disaster in history, and it was caused entirely by the mismanagement of water.

The Ogallala Aquifer, stretching beneath the Great Plains of the United States from South Dakota to Texas, irrigates roughly thirty per cent of all American agricultural land. It is being depleted at a rate three to ten times faster than natural recharge. In parts of western Kansas and the Texas Panhandle, the water table has dropped by more than forty-five metres since intensive pumping began in the 1950s. At current extraction rates, large sections of the aquifer will be effectively exhausted within thirty to fifty years (Steward et al., 2013). The Ogallala took millions of years to fill. It is being emptied in a single human lifetime.

India pumps more groundwater than any other country on Earth — an estimated 251 cubic kilometres per year, more than the United States and China combined (FAO AQUASTAT, 2023). Over sixty per cent of Indian irrigated agriculture depends on groundwater, and water tables are falling across the Indo-Gangetic Plain at rates that threaten the food security of hundreds of millions of people. In 2019, a government think tank — NITI Aayog — warned that twenty-one Indian cities, including Delhi, Bangalore, and Hyderabad, would run out of groundwater by 2025. That prediction was optimistic. Chennai, a city of ten million, effectively ran dry in June 2019, with reservoirs falling to 0.1 per cent of capacity. Water was delivered by train — literally, a fifty-wagon train carrying 2.5 million litres from a town two hundred kilometres away.

São Paulo, a metropolitan area of twenty-two million people, came within twenty days of emptying its main reservoir system in 2015. Mexico City, built on a drained lakebed, is sinking at up to thirty centimetres per year as the aquifer beneath it is pumped dry, and forty per cent of its water is lost to leaking pipes. Jakarta is sinking so fast from groundwater extraction that the Indonesian government is building an entirely new capital city on Borneo (WRI Aqueduct, 2023).

The World Resources Institute’s Aqueduct data identifies twenty-five countries — home to one quarter of the global population — as facing “extremely high” water stress, meaning they withdraw more than eighty per cent of their renewable supply every year (WRI Aqueduct, 2023). Among them are some of the world’s most volatile regions: the Middle East, North Africa, and South Asia.

We have not transcended the ancient dependency on water. We have merely borrowed against the future — pumping aquifers that took millennia to fill, draining rivers that sustained ecosystems for millions of years, and building cities of tens of millions in places where the natural water supply supports a fraction of that population. The bill is coming due.

Water Wars and Water Politics

The English word “rival” derives from the Latin rivalis — one who uses the same stream. The etymology is not a coincidence. Competition over shared water sources is as old as organised human society, and it is intensifying.

The Nile is the most consequential current flashpoint. Egypt — a nation of 105 million people in which ninety-seven per cent of the land is desert — depends on the Nile for virtually all of its freshwater. Under the 1959 Nile Waters Agreement (which Sudan co-signed but upstream nations did not), Egypt claims 55.5 billion cubic metres of the river’s annual flow. Ethiopia, the source of roughly eighty-five per cent of the Nile’s water via the Blue Nile, was not a party to that agreement and does not recognise it. In 2011, Ethiopia began constructing the Grand Ethiopian Renaissance Dam (GERD) — a 6,000-megawatt hydroelectric facility on the Blue Nile that, when full, will hold 74 billion cubic metres of water. Egypt views the dam as an existential threat. Ethiopian officials view it as an existential necessity — thirty-five million Ethiopians lack electricity, and the dam could power the nation’s industrialisation (Solomon, 2010).

Negotiations have stalled repeatedly. Egyptian officials have made barely veiled threats. In 2013, a televised meeting of Egyptian politicians — broadcast accidentally when participants did not realise the cameras were live — included open discussion of sabotaging the dam or supporting Ethiopian rebel groups. The dam is now largely complete and filling, but the long-term allocation of Nile water remains unresolved. Egypt’s population is projected to reach 160 million by 2050. Ethiopia’s will reach 205 million. The river is not getting any larger.

The Indus Waters Treaty between India and Pakistan, brokered by the World Bank in 1960, has survived three wars, a nuclear standoff, and decades of hostility. It allocated the three eastern tributaries of the Indus to India and the three western tributaries to Pakistan. It is widely regarded as one of the most successful water-sharing agreements in history. But it is under strain. India is constructing hydroelectric projects on the western tributaries that Pakistan alleges violate the treaty. Pakistan depends on the Indus basin for ninety per cent of its food production. In 2016, following a terrorist attack attributed to Pakistan-based militants, Indian Prime Minister Narendra Modi declared that “blood and water cannot flow together” — a pointed suggestion that India might reconsider the treaty. For a nuclear-armed state whose agriculture depends entirely on a river controlled upstream by its rival, this is not idle rhetoric.

Turkey’s Southeastern Anatolia Project (GAP) — a network of twenty-two dams and nineteen hydroelectric plants on the Tigris and Euphrates — has reduced downstream water flow to Iraq and Syria by an estimated forty to eighty per cent, depending on the season (Beaumont, 1998). Iraq, which once possessed some of the most productive agricultural land in the Middle East, has lost vast stretches of farmland to drought and salinisation. Southern Iraq’s Mesopotamian Marshes — the largest wetland ecosystem in the Middle East, once covering twenty thousand square kilometres — were reduced to ten per cent of their former area by a combination of upstream damming and Saddam Hussein’s deliberate draining programme (UNEP, 2001). The cradle of civilisation is drying out, and the water that once flowed through it is generating electricity in Turkey.

No two nations have gone to war solely over water — yet. But as Steven Solomon argues in Water: The Epic Struggle for Wealth, Power, and Civilization (2010), water scarcity has been a contributing factor in nearly every major regional conflict of the past century, from the Syrian civil war (preceded by the worst drought in the Fertile Crescent in nine hundred years) to the ongoing instability in the Sahel. Water does not cause wars by itself. But it makes every other source of conflict — ethnic, religious, economic, territorial — significantly worse.

The Future — Desalination, Conservation, or Conflict?

There is a hopeful version of the future, and it has a name: Israel.

Israel is, by any natural measure, a country that should not have enough water. It sits on the edge of the Negev Desert, receives an average of five hundred millimetres of annual rainfall (compared to London’s six hundred), and has a population that has grown tenfold since 1948. By every historical precedent, it should be in permanent water crisis.

Instead, Israel now produces more freshwater than it consumes. The country operates five major desalination plants along the Mediterranean coast that collectively produce roughly 585 million cubic metres of freshwater per year — approximately eighty per cent of domestic water consumption (Tal, 2016). It recycles eighty-seven per cent of its wastewater for agricultural use, the highest rate in the world. Its drip irrigation technology, pioneered by Netafim in the 1960s, reduces agricultural water use by up to seventy per cent compared to flood irrigation. Israel is the only country in the world where the desert is shrinking.

Singapore has followed a similar path. A city-state with virtually no natural freshwater sources, Singapore developed “NEWater” — ultra-purified recycled wastewater that meets forty per cent of the nation’s demand and is projected to reach fifty-five per cent by 2060. Singapore also operates one of Asia’s largest desalination plants (PUB, 2023). It treats water security as a matter of national survival, which, for a country that once depended entirely on water imported from Malaysia, it is.

Saudi Arabia is the world’s largest producer of desalinated water, operating thirty desalination facilities that supply roughly fifty per cent of the kingdom’s drinking water. But Saudi desalination is almost entirely powered by oil and natural gas, consuming an estimated 1.5 million barrels of oil equivalent per day (World Bank, 2022). This exposes the central paradox of desalination: it solves the water problem by creating an energy problem. Desalinated water costs three to ten times more than conventional water sources, and every litre requires between three and ten kilowatt-hours of energy, depending on the technology (UN World Water Development Report, 2024). For wealthy nations with abundant energy, this is manageable. For the developing world — where the water crisis is most acute — desalination remains largely out of reach.

The optimistic case is that technology solves it. Solar-powered desalination is becoming viable. Membrane technologies are reducing energy costs. Smart irrigation and precision agriculture are cutting water waste. Atmospheric water generation — pulling moisture from the air — is moving from laboratory curiosity to commercial deployment. Israel has demonstrated that a nation can, with sufficient investment and political will, engineer its way out of water scarcity.

The pessimistic case is that most of the world is not Israel. The countries facing the most severe water crises — Pakistan, India, Egypt, Iraq, Ethiopia, the Sahel nations — are precisely those least able to afford the multi-billion-dollar desalination infrastructure, the energy to run it, and the institutional capacity to manage it. They are also, not coincidentally, the countries with the fastest-growing populations. Sub-Saharan Africa’s population is projected to double by 2050, reaching 2.1 billion — in a region where forty per cent of the population already lacks access to basic drinking water (UN World Water Development Report, 2024).

The historical record offers no reassurance. Every civilisation that ran out of water collapsed. The Akkadians did not find a technological solution. The Maya did not invent desalination. The Khmer did not adapt in time. The question is whether the modern world — with its science, its capital, and its awareness of the risk — will be the first civilisation to break the pattern, or merely the largest to confirm it.

The water is still flowing. But the taps are tightening. And history is rather clear about what happens next.