An Anthropogenic Tsunami in an Unusual Setting

The Vaiont disaster is particularly interesting as an example of a tsunami-like event in which human actions played a significant part, although it should be noted that the exact sequence of events is still the subject of litigation in Italy and some aspects are still not fully clear. The following represents an interpretation of events proposed by engineers from the US Army Corps of Engineering, who were involved in the post-disaster enquiries as impartial outside experts.

The Vaiont dam, at the time the tallest concrete arch dam in the world at some 260 m high, was built at the entrance to a very narrow and deep valley tributary to the Piave river valley in the Tyrolean Alps of NE Italy. The purpose of the dam was to impound water for a hydroelectric power station. As the valley filled with water after completion of the dam in 1960, an ancient landslide on its upper southern side, adjacent to the dam, began to move again, in episodes of slow creeping movement. It was later established that this was due to groundwater, unable to escape into the floor of the now - flooded valley, saturating a layer of clay within the rocks beneath it. This accumulation of water high in the slope was most marked during periods of heavy rain, although the association of heavy rainfall and creep went unnoticed. However, the presence of the impermeable clay layer in the bedrock was also not recognized at the time and it was assumed that the movement was due to local saturation of the rocks below the level of water in the reservoir, at the toe of the creeping landslide, rather than accumulation of water pressure in the entire mountainside. It was therefore proposed to regulate the movement of the landslide, and thus allow it to settle to a new equilibrium, by lowering the level of water in the lake when an episode of creep was in progress, until the toe of the landslide was no longer saturated and the creeping stopped The reservoir was then allowed to refill and the drainage cycle repeated whenever creeping movement occurred again.

On 8 October, after ten days of heavy rain, the dam engineers attempted to arrest the creeping motion of the landslide by partially draining the reservoir again. However, because lowering the level of the lake immediately removed a stabilizing force, that of the weight of water in the lake pushing against the toe of the landslides, whilst not immediately draining water from the upper part of the mountain because of the barrier to flow represented by the clay layer, the net effect of their action was actually to destabilize the landslide rather than stabilize it. The movement accelerated through 9 October, as the engineers continued to attempt to arrest movement of the slide by lowering the reservoir level (although in fact their efforts merely served to remove the arresting buttress provided by the weight of water on the toe of the slide). Late that evening, the moving rock mass, with a volume of 240 million cubic metres, failed catastrophically and slid into the reservoir within about 30 seconds, at an average velocity of 20-30 metres per second, or up to 110 km/h.

Remarkably, the dam withstood the weight of rock piling up behind it and it remains largely intact to this day. However, the landslide displaced as much as 50% of the water within the lake: a tsunami wave traveled east up the reservoir, destroying two villages, whilst to the west tsunami-like surges of water swept up the valley sides to heights of 300 m above the reservoir level and over the top of the dam. A flood wave 70 m high swept over the town of Longarone and both up and down the Piave valley. A total of 2600 people were killed; economic losses have never been definitively assessed, as a result of the litigation mentioned above.