CountriesKnowledge Hubs
Data & Tools
Region AtlasMajor ProjectsPPP Pipeline & AwardsLive tendersWhere to find live tendersTariffs & cost benchmarksOrganization DirectoryMarketsIntelligenceStoriesCalendarAbout
The old city of Sanaa, Yemen, seen across its tower houses in daylight

MENA Water Review · Visual story

The Slow Death of Water

Yemen was once the most accomplished dryland water civilisation on earth. Then it learned to drill, and its capital began running out of water long before the war arrived.

Running out of water is not an event.

For two decades, Sanaa has held the same distinction in the water-sector literature: it is the first modern capital likely to exhaust its water. The predicted dates have passed, including 2017, yet the city is still drinking. The forecasts were not wrong about the resource.

No city of several million wakes one morning to find its aquifer empty. Instead, the water table falls, making water more expensive, and access is rationed by price rather than through the pipe network. Households that once relied on a municipal connection buy from tankers. Tanker prices rise. Families further down the income scale buy less water and use water they would not otherwise have chosen. A city does not so much run out of water as sort itself into those who can still afford it and those who cannot.

Absolute scarcity threshold: 500 m3 per personAbsolute scarcity threshold500 m3 per personYemen, 2022: about 55 m3 per personYemen, 2022about 55 m3 per person

Internal renewable water per person per year

The Falkenmark indicator treats 500 cubic metres per person per year as absolute scarcity. Yemen has roughly a ninth of that, among the lowest figures recorded anywhere.

~55 m³of internal renewable water per person per year, 2022
4–6 mthe estimated annual fall in groundwater levels beneath Sanaa
~13,000wells drilled without regulation in the Sanaa basin
2×+groundwater overdraft against the rate of recharge
Stone agricultural terraces on the slopes of the Haraz Mountains, Yemen
The Haraz Mountains. Hand-built terracing on this scale is water infrastructure as much as it is farmland: it slows the rain, forces it to soak in rather than run off, and holds the soil in place.

Before the pump, none of these systems could be over-drawn.

The usual account makes Yemen seem simply careless with its water. The opposite is the case. On the mountainsides, farmers built stone terraces in their thousands. Some highland systems were first laid in the Bronze Age and are still in use. In the wadis, they practised spate irrigation, diverting the brief, violent floods that follow a storm across fields prepared to receive them.

All these systems share a design principle: each can use only water that has recently fallen from the sky. A flood cannot be over-drawn. The rules that governed them, upstream before downstream and ancient rights before new ones, were enforced by communities that all drew on the same visible source. They could see at once when a neighbour took more than his share.

Bare slopeTerraced slopeThe same rain falls on bothRuns off, taking the soil with itHeld, and soaked inRecharges the spring below
1 of 3

Two hillsides under the same storm. The difference between them is entirely a matter of what was built on the slope.

On the bare slope the water runs off within minutes and carries the soil with it. On the terraced slope each wall holds it long enough to soak in.

Water held on the terraces recharges the springs below, which is what the villages drank from. A terrace that is not repaired collapses, and the soil behind it washes out in the next storm.

The well was, for each farmer, an entirely rational decision.

In the 1970s and 1980s, Yemeni men worked in the Gulf in very large numbers and sent billions of dollars home. Some of that money bought a diesel pump and a borehole. This is the hinge of the whole account, and it is worth stating plainly: for an individual farmer, the well was an unambiguously good decision. It freed him from waiting for rain that might not come, allowed him to grow a cash crop rather than subsistence grain, and made his family more secure.

The resource he drew on, however, was shared and invisible. The customary rules worked precisely because everyone could see the same source, so they had nothing to say about it. What followed is the most familiar sequence in groundwater management and one of the hardest to arrest. As the water table falls, shallow wells fail. Those who can afford to drill deeper keep their water; those who cannot lose theirs. Depth becomes a proxy for wealth. Each new deep well lowers the table further, making it more necessary, rather than less, to drill deeper.

0 m200 m400 m600 m800 m1000 mGround levelEarly 1970s: under 30 mthe water table, within reach of a shallow well1995: below 150 mthe water table, after two decades of pumping2010: wells 100 to 400 mdrilling depth, not the table2010: and 300 to 500 mdrilling depth into the volcanic aquiferThe deepest: 1,000 ma few wells, as reported in 2010Experts put the fall in groundwater levels at four to six metres a year
1 of 5

The water table beneath the capital stood at under 30 metres, within reach of a shallow hand-dug well.

Two decades of pumping had taken it below 150 metres.

New wells were being drilled 100 to 400 metres into the volcanic aquifer. From here the figures are drilling depths rather than measurements of the table itself.

Others were bored to between 300 and 500 metres, each one lowering the table a little further for everyone around it.

A few reached 1,000 metres. Around Sanaa the basin yields about 200 million cubic metres a year against an input of roughly 50 million.

The first two figures are water-table depths; the rest are drilling depths reported in 2010 by the World Bank-funded Sanaa Basin Water Management Project. Individual wells vary widely.

Qat pays every week of the year, and grows where water must be lifted.

No account of Yemeni water proceeds far without qat, the mild stimulant leaf chewed socially by a large share of the adult population. From outside, it is usually described as a national vice. It is more useful to understand qat as the crop that makes most sense to grow. It is perennial, so it need not be replanted. It is harvested continuously rather than once a season, so it yields income year round. It is sold fresh for cash on the same day, and it faces no imported competition of the kind that undercuts grain.

Qat is also thirsty, and it is grown where water has to be lifted. Estimates put qat at about 30 per cent of groundwater use, and in 2004 it covered nearly a fifth of the harvested irrigated area. In this sense, Yemen’s water problem shows what happens when the profitable use of a free resource and its sustainable use point in opposite directions.

Agriculture: about 90% of withdrawalsAgricultureabout 90% of withdrawalsQat alone: about 30% of groundwater useQat aloneabout 30% of groundwater use

Where Yemen’s water goes

Withdrawals for 2000 from FAO AQUASTAT; the qat share from the Sanaa Center. The two are measured on different bases, so the second bar is shown as a share of groundwater use rather than of the first bar.

The remains of the Great Dam of Marib in eastern Yemen
Marib. The Great Dam held the floods of a desert wadi for well over a thousand years and made the kingdom of Saba wealthy enough to be remembered as Sheba. It required constant skilled maintenance, and when that faltered it broke.

Yemen has known for fourteen centuries what happens when the institutions fail.

East of Sanaa stand the remains of the Great Dam of Marib. Inscriptions credit Sabaean rulers with building it across Wadi Adhanah in the seventh century BCE. It was breached and repaired in 450, and again in 542. Around 570 CE, it broke for the last time. The irrigated land reverted to desert, and many of its people moved away, an event the Qur’an itself recalls.

The water did not disappear first. The capacity to manage it did.

  • 7th c. BCE

    The Great Dam of Marib

    Sabaean rulers dam Wadi Adhanah. Its branching canals water the fields of a kingdom wealthy enough to dominate South Arabia.

  • c. 570 CE

    The dam fails for the last time

    After breaches repaired in 450 and 542, the structure gives way for good. The irrigated land reverts to desert.

  • Bronze Age on

    Terraces and spate irrigation

    Mountain terracing and wadi flood-spreading, governed by detailed customary water law, sustain rural populations on rainfall and floods alone.

  • 1970s

    Remittances and rigs

    Money sent home from the Gulf, heavily subsidised diesel and new drilling rigs put a private well within reach of an ordinary farmer.

  • 1980s–90s

    The drilling boom

    Groundwater use outruns customary control. The area under qat more than triples in twenty-five years.

  • 2000s

    The warnings become explicit

    The Sanaa basin yields about 200 million cubic metres a year against an input of roughly 50 million, and experts warn the capital could run short by 2017.

  • 2015

    War

    Fuel scarcity cripples pumping and power. Across sixteen cities, 38 per cent of water and sanitation assets are damaged, and civil servants’ salaries fall years into arrears.

  • 2016–21

    Cholera

    The largest cholera outbreak of its time: more than 2.5 million suspected cases.

  • 2019–26

    The need does not recede

    The UN counted 17.8 million people without adequate water, sanitation and hygiene in 2019. In June 2026, 14.4 million still required humanitarian assistance with them.

The war did not cause the crisis. It removed every remaining buffer.

Yemen’s water crisis was already decades old when the war began in 2015. Water is pumped in Yemen, and pumping requires fuel. Fuel scarcity and the collapse of the power grid left boreholes, treatment works and sewage plants struggling to operate. Networks were damaged, some directly, and civil servants went unpaid for years. Households fell back on tankers of uncertain quality, unprotected wells, and whatever could be collected.

The consequence came in October 2016, when one of the largest cholera outbreaks on record arrived. Cholera is not an exotic disease. It spreads through food or water contaminated with the bacterium, and safe water and basic sanitation prevent it. That is precisely the infrastructure Yemen had been building but could no longer operate.

2019: 17.8 million people201917.8 million peopleJune 2026: 14.4 million peopleJune 202614.4 million people

Without adequate water, sanitation and hygiene

Counted on different bases seven years apart: the 2019 figure is people without adequate access, the 2026 figure people requiring humanitarian assistance. Both are drawn from UN assessments and are shown together to indicate scale rather than a trend.

Solar pumping removed the last thing limiting extraction.

When fuel for pumps became scarce and expensive after 2015, Yemeni farmers did something genuinely impressive: they turned to the sun. Markets for panels boomed, and solar pumping spread across the farms that could afford it. By every ordinary measure, this is good news. It is clean, resilient, and works when the fuel convoys do not. It has kept water flowing through a war.

FuelPanelDiesel: every hour costs money, so there is a reason to stopSolar: once installed it costs nothing to run, so it runs

A diesel pump meters itself: every hour of running costs money, so every farmer has a reason to stop. A solar pump, once installed, costs essentially nothing to operate, so the rational thing is to run it whenever the sun is up.

Without the discipline of a fuel bill, solar pumping can exceed what diesel ever extracted, and local authorities do not regulate it. Yemen has therefore replaced a dirty constraint with a clean absence of one.

It would be dishonest to end with an optimistic flourish. Groundwater is being drawn at more than twice the rate at which it refills, in a country with no money, public institutions that can no longer deliver services, and an unfinished war. But the picture is not uniform. Terrace rehabilitation restores rainwater capture and is labour-intensive, which, in a collapsed economy, is a feature rather than a cost. Spate irrigation still functions in the wadis. Local water user associations have taken over the running of schemes that the state no longer reaches.

The uncomfortable conclusion is that Yemen’s water problem is not, at root, hydrological. The country knew how to manage scarce water and did so for thousands of years, and the physical evidence of that knowledge still stands on every mountainside. What broke was the link between the rules and the resource, when a technology arrived that allowed individuals to reach water that no rule covered.

Images: old city of Sanaa © Hamza Shiban (CC BY-SA 3.0); Haraz Mountains terraces © Rod Waddington (CC BY-SA 2.0); Great Dam of Marib © H. Grobe (CC BY-SA 3.0). Via Wikimedia Commons.

Sources

22 references

Groundwater figures for the Sanaa basin are estimates: the aquifer is not comprehensively monitored and the well count is itself an approximation. Depths after 1995 are drilling depths rather than measurements of the water table.

  1. IRIN (now The New Humanitarian), “Capital city faces 2017 water crunch”, 23 March 2010.
  2. World Bank (2021), project appraisal document, Second Integrated Urban Services Emergency Project.
  3. World Bank, Yemen country overview, accessed September 2026.
  4. World Bank, World Development Indicators, renewable internal freshwater resources per capita (cubic metres): Yemen, 2022 value.
  5. White, C. (2012), “Understanding water scarcity: definitions and measurements”, Global Water Forum, 7 May 2012.
  6. World Bank and GW-MATE (2003), Yemen: Rationalizing Groundwater Resource Utilization in the Sanaa Basin.
  7. FAO AQUASTAT, country profile: Yemen (2008 edition, archived).
  8. UNDP (2022), A holistic approach to addressing water resources challenges in Yemen.
  9. Pietsch, D. and Mabit, L. (2012), “Terrace soils in the Yemen Highlands: using physical, chemical and radiometric data to assess their suitability for agriculture and their vulnerability to degradation”, Geoderma, doi:10.1016/j.geoderma.2012.03.027; abstract via edoc, University of Basel.
  10. Stewart, R.T. (1978), “A dam at Marib”, Saudi Aramco World 29(2), March/April 1978.
  11. UNESCO World Heritage Centre, Landmarks of the Ancient Kingdom of Saba, Marib (archived).
  12. World Bank (2020), Yemen Dynamic Needs Assessment: Phase 3 (2020 Update).
  13. WHO EMRO, Cholera situation in Yemen, April 2021, via ReliefWeb (PDF).
  14. Heffez, A. (2013), “How Yemen chewed itself dry”, The Washington Institute for Near East Policy, 23 July 2013.
  15. Sanaa Center, “Qat expansion in Yemen’s war economy: water depletion and poisonous pesticides”, 2026.
  16. World Health Organization, cholera fact sheet, accessed September 2026.
  17. World Health Organization, Weekly Epidemiological Record, 21 September 2018, vol. 93, no. 38 (global cholera report for 2017), via ReliefWeb.
  18. UN OCHA (2019), Yemen: Humanitarian Needs Overview 2019.
  19. Global WASH Cluster, UNICEF and ECHO, WASH Insecurity Analysis (WIA): Yemen snapshot, June 2026, posted 24 August 2026.
  20. Sanaa Center, “Solar-powered irrigation in Yemen: opportunities, challenges and policies”.
  21. Conflict and Environment Observatory, “Groundwater depletion clouds Yemen’s solar revolution”, April 2021.
  22. United Nations in Yemen, “Being the change in Yemen: improving integrated water resources management for food security”, 23 March 2023.

Every figure and claim in this story rests on the references above. The stories list them here rather than marking them in the text; the country reports and hubs keep their inline markers.