The eccentric engineer: the catastrophic outcome of trying to reengineer a mountain

Rarely is physical engineering seen more spectacularly than in hydroelectric dams. Apparently taming whole wildernesses is the apogee of the discipline for many. Yet structural engineering is at least in part about knowing your limitations, as the workers of Vajont found out to their cost.

The dam at Vajont had been conceived in the 1920s as part of Mussolini’s plan to industrialise the north of Italy, but was only authorised in 1943, after Il Duce’s ignominious fall. It was his former finance minister, Giuseppe Volpi di Misurata, whose company Società Adriatica di Elettricità (SADE) then got the job of controlling and harnessing the Piave, Mae and Boite rivers.

In the early 1950s, engineering surveys apparently confirmed the suitability of the site, and construction began in 1957 on what was to be, at 263m high, the world’s tallest thin-arch dam.

Just two years later, fractures were noticed on the new road built on the side of Mount Toc that bordered the dam’s reservoir. Engineering studies suggested the entire side of the mountain might be unstable. Yet work went on and in February 1960 the company was authorised to begin filling the basin.

During that summer, several minor landslides and earth movements were recorded as water began to press on the sides of the reservoir. However, media reports were ignored as being written by communist agitators. In October 1960, a huge fissure opened up, 2km long and defining an area about 1,700m long and 1,000m wide on the left bank of the dam. Then on 4 November some 800,000m3 of material collapsed into the lake.

This could not be ignored and SADE immediately drained 50 metres of water to relieve pressure. Further engineering studies suggested that more landslides could be expected, potentially dividing the lake in two, but construction of an artificial gallery in the basin in front of Mount Toc would allow continued electricity generation even with this blockage. With SADE’s concerns about the economic viability of the project allayed, and the gallery built, filling resumed.

By the spring of 1962, the water depth had reached 215m and there were local reports of magnitude 4-5 earthquakes in the vicinity. Despite this, the company was authorised to fill the reservoir to its full depth.

By that summer SADE’s own engineers, who had built models of the entire dam system, were warning that a sudden landslide might create a wave large enough to overtop the dam and crash down into the valley. As a result, the water level was lowered to 25m below the dam top.

With the probability of a landslide now clear, a dangerous idea was forming among SADE’s engineers – as so massive a landslide would be impossible to engineer away, perhaps it might be ‘controlled’ by adjusting water levels in the reservoir.

A further reduction in water level seemed to suggest they might be right. As the level was reduced to 185m in November 1962, creep substantially reduced and by the following April had stopped altogether, spurring the engineers to begin refilling. Over the next month, the reservoir was rapidly filled to 237m and then to 245m by early September, at which time the land-creep had reached 3.5cm a day in places. By 9 October, the increased rate worried engineers enough for a further drawing down of water levels. Yet this time the creep increased, reaching 20cm a day.

That very evening, the entire side of the mountain failed at a rate of 30m/s, filling the gorge to a depth of 400m. In just 45 seconds a 260m-high wave was pushed up the opposite bank destroying a village high above the dam. Far worse was to come. The landslide displaced around 30 million cubic metres of water, which now overtopped the dam with a staggering 245m-high mega tsunami, crashing down onto the villages of Longarone, Pirago, Villanova, Rivalta and Fae half a kilometre below. The shock wave just from the displaced air was estimated to be twice that of the Hiroshima atomic blast, atomising many of the inhabitants below whose remains were never found. The official death toll was 1,917, but it is thought perhaps 2,500 died with 350 families being entirely wiped out.

Remarkably, the dam survived. It was, as had been planned, an engineering marvel. Little blame fell on the company or the government, with the newspaper of the party in power calling the event an act of God. Engineers might know better.