The Lake That Runs Both Ways

The Oued Tinja is a short channel between Lake Ichkeul and Lake Bizerte, and for part of every year the water in it moves the wrong way. In winter the six wadis that feed Ichkeul — the Joumine, the Sejnane, the Ghezala, the Melah, the Tine and the Douimis — come down off the hills and push a lake's worth of fresh water out through the Tinja towards the sea. In summer the rain stops, evaporation drops the lake below sea level, and the Mediterranean comes back up the same channel through Lake Bizerte, salt and slow. There are very few places where the direction of a river is a seasonal variable. This is one, and it is not a curiosity. It is the mechanism the whole place runs on.

The winter flush keeps salinity low enough for beds of pondweed, Potamogeton pectinatus, to cover the lake bed, and for bulrush marshes to stand around the edges. The summer reversal keeps the system brackish rather than fresh, which is what stops it silting into an ordinary reedbed. Together they produce a wintering ground that has held numbers of waterfowl not matched anywhere else in North Africa — greylag geese in tens of thousands, coot, wigeon and pochard in hundreds of thousands. UNESCO inscribed the park in 1980. Sixteen years later it put it on the List of World Heritage in Danger.

What the dams took

Tunisia dammed the north because it had to. Almost all of the country's reliable surface water falls in these hills, and almost all of the people and the irrigated farmland are somewhere else. The Joumine dam closed in 1983, the Ghezala in 1984, the Sejnane in 1994, and between them they captured the large majority of the fresh inflow that had been arriving at Ichkeul each winter, sending it instead to fields and to the taps of Bizerte and Tunis. Nobody set out to damage the lake. The lake was simply downstream of a national water plan and was not, at the time, a line in it.

What followed was not a drying-out. It was the loss of one half of a rhythm. With little fresh water arriving, the summer reversal ran longer and pushed further, and a lake whose salinity had swung seasonally around a low annual average began spending whole years at something close to sea strength. The pondweed beds, which had covered thousands of hectares, thinned as each dam closed and went at the end — the Sejnane, the last of the three, only shut in 1994, so the worst of it is a 1990s story rather than a 1980s one — and by the turn of the century they had effectively gone. Bulrush was replaced by salt-tolerant species the geese will not eat. Greylag numbers fell from tens of thousands to a few hundred, and the wintering count, which had run into the hundreds of thousands, bottomed out at somewhere around eleven thousand in 2003. A lake had been converted, without anyone deciding to, into an arm of the sea.

The dams did not take the lake's water. They took the winter half of a rhythm and left the summer half running unopposed, which is a subtler thing to do and a harder one to notice in time.

The gate

The answer Tunisia arrived at has two parts, and both are unglamorous. The first is a sluice across the Oued Tinja, agreed in 1993 and in place by 1996 — a gate that can be shut to stop the sea coming up the channel in summer and opened to let the lake drain in winter. The second is the part that matters more: a commitment to release fresh water from the dams for the lake itself, on the order of 80 to 120 million cubic metres a year, water that would otherwise irrigate crops or fill a reservoir for a city. Ichkeul was given, in effect, a water right in a country that has never had enough water to go round.

Neither half works alone, and the sluice in particular is a blunter instrument than it looks. It has to be closed at the right moment, and closing it means accepting that the lake will not drain — which is only tolerable if there is fresh water coming in to replace what evaporates. In the dry winters of 1999 and 2000 there was not enough inflow to justify shutting the gate, so it stayed open and the sea came in anyway. The gate is a lever on a system; it is not a lid.

Off the list

Then it rained. The winter of 2004 and the one after it were good ones, the dam releases went out on top of them, and the lake responded faster than anyone expected: salinity dropped back into the range the pondweed can live in, the pondweed came back, the fish and the eels came back, and the wintering count climbed off that 2003 floor and into six figures again within a couple of seasons. The World Heritage Committee took Ichkeul off the danger list in 2006, one of the short roster of properties that have ever come back off it. It is worth being precise about what that proves. The rain did a great deal of the work and no management regime can produce rain; what the sluice and the releases did was make the rain count, holding the fresh water in when it arrived instead of letting it run out past an open gate while the sea pushed in behind it. That is a real result and a modest one, and the Tunisian hydrologists and park staff who argued for it spent a decade doing so against every competing claim on the same water.

What it costs to keep

Ichkeul is now an actively managed lake. Its fresh water arrives when a person decides to open a valve on a dam, and its salt water is excluded when a person decides to close a gate at Tinja. That is a permanent operating commitment, and it is being made by a country in the middle of a water emergency. On 31 March 2023 SONEDE, the national water utility, introduced quotas and began cutting the mains supply overnight, roughly nine in the evening until four in the morning. Dam storage was down to about a billion cubic metres, some 30 per cent of capacity; Sidi Salem, the largest reservoir in the country, was at around 16 per cent. Using drinking water to irrigate a field, wash a car or hose a pavement was made an offence carrying fines and, in the worst cases, a prison term. The drought that forced it has eased since; the arithmetic under it has not. Tunisia has well under 500 cubic metres of renewable fresh water per person per year, which is the line hydrologists call absolute scarcity, and it is on the wrong side of that line in a normal year, not only a bad one.

Set the two facts beside each other and the shape of the decision is clear. In a year when households in Tunis have no water between nine and four, somewhere between 80 and 120 million cubic metres is released down the Joumine and the Sejnane so that pondweed can grow on a lake bed and geese can winter on it. There is no version of that which is automatic. It has to be re-argued against agriculture and against the cities every season, and it is won on grounds — a World Heritage listing, a Ramsar designation, a biosphere reserve, the standing of the people making the case — rather than on hydrology, because on hydrology alone the lake loses every time. A visitor sees almost none of it, which is the reason to point it out: Ichkeul in February is a shallow sheet of water under a limestone ridge, a marsh track, an eco-museum on the hillside explaining the salinity curve, a disused hot spring and, if the winter has been kind, more birds than take in at once. Twenty minutes away there is a concrete sluice across a narrow channel with a road over the top of it and nothing whatever to look at. Stand on it for a minute anyway. Everything upstream of it is a policy that gets renewed, or does not.