The Water Under Marrakech
Marrakech was built on 700 kilometres of khettara — hand-dug tunnels carrying Atlas groundwater to the city by gravity, incapable of over-extracting because they had no pumps. The motor well replaced them in a generation, and the aquifer has been falling ever since.
Under the palm groves north of Marrakech there is a line of shallow circular pits, thirty or forty metres apart, running toward the Atlas. From the air they look like a dotted line drawn across the plain. They are the access shafts of a khettara — an underground canal, dug by hand at a gradient of a few metres per kilometre, carrying groundwater from the foot of the mountains to the city by gravity alone.
Marrakech was founded in 1070 and it exists because of these. The Almoravids built the first network; by the twentieth century the city was fed by something like 700 kilometres of khettara, some of them running twenty kilometres or more. Almost all of them are now dry.
How the system worked
The engineering is elegant and unforgiving. A khettara taps the water table where it is high, near the mountains, and runs a tunnel toward the city at a gradient shallower than the natural slope of the land, so that the tunnel gradually rises relative to the water table until it emerges at the surface. No pumps. No power. The vertical shafts every few dozen metres are for digging and for maintenance, and they are the reason the surface line is visible at all.
Critically, the system is self-limiting. A khettara can only ever deliver what the water table will give it by gravity. If the table drops, the flow drops. It cannot over-extract, because it has no mechanism for doing so. A thousand years of Marrakech's agriculture and its gardens — the Agdal, the Menara, the palm grove — were built inside that constraint.
The maintenance was collective. Khettara belonged to groups of users who shared the water on a rotation measured in time rather than volume, and who were jointly responsible for keeping the tunnel clear. It is dangerous work; collapses kill people. The institution that organised it was as important as the tunnel.
A khettara cannot over-extract. It delivers what gravity gives it. The tube well that replaced it has no such limit, which is the entire problem.
What replaced it
The motor pump arrived in the middle of the twentieth century and changed the arithmetic completely. A tube well reaches water a khettara cannot, delivers it on demand rather than on rotation, and requires no collective institution to maintain. Within a few decades the khettara network was largely abandoned, not because anyone decided to abandon it but because each individual farmer who drilled a well was making an obviously sensible decision.
The aggregate result was a falling water table, which dried the khettara that were still running, which pushed the remaining users onto wells. The Haouz plain aquifer beneath Marrakech has dropped by tens of metres in places since the 1970s. Morocco has been in drought for most of the years since 2018, reservoir levels have hit historic lows, and the Oum Er-Rbia and Tensift basins have been under emergency management.
Golf courses and the argument about them
Marrakech has a dozen or so golf courses and they are the lightning rod. A course in an arid climate uses a very large amount of water; a dozen of them next to a city under water restriction is an easy image and a genuinely provocative one. The courses respond that most now irrigate with treated wastewater, which is true of several of them and is a real improvement.
The honest numbers are less satisfying for everybody. Agriculture uses the overwhelming majority of Moroccan water — around 85% — and within that, export crops are the pressure point. Watermelon in the Zagora region, avocado and citrus in the Souss, and soft fruit for European supermarkets all consume aquifer water on a scale that dwarfs the golf courses. Zagora had water protests in 2017 while watermelons grown nearby went to market. The government has since restricted watermelon planting in the worst-affected provinces.
Tourism's share is real but smaller, and the useful version of the argument is not about whether a hotel pool is defensible in the abstract. It is about whether the water it uses is metered, priced and reported, and in a growing number of cases it now is.
Whether the khettara come back
There are restoration projects. Several khettara in the Haouz and in the southern oases around Tafilalt and Skoura have been cleared and returned to service, usually with development funding and usually as much for heritage and tourism as for water supply. Where the aquifer still has enough head, they work exactly as designed.
Where it does not, they cannot, and no amount of restoring the tunnel changes that. A khettara is a device for reading the water table, and a dry one is an accurate reading. What the restorations have arguably recovered is the other half of the system — the collective institution for allocating a shared and limited resource, which is the part a tube well made unnecessary and which everyone now needs again.
Where to see them
The Musée de l'Eau et de la Civilisation Marocaine, on the road toward the Ourika valley, does the hydraulic history properly and has a walkable khettara section. In the palm grove north of Marrakech the shaft lines are visible from the road if you know what you are looking at.
The better versions are in the south. The oases of the Draa and the Tafilalt around Rissani still run khettara that feed date palms, and the Skoura palm grove is a working example of the whole system — canals, rotation, and gardens under palms that only exist because of it. Ask at a kasbah guesthouse; the family will almost certainly hold a share of a water rotation and can explain how the timing is divided, which is the part no museum conveys.
Kwame edits ExplWorld’s Africa coverage and drives the continent’s worst roads personally to keep the guides honest. Formerly of the BBC’s Lagos bureau, he has crossed the Sahel five times and maintains the magazine’s overland-border database himself.