Drone models reveal where the Pakri cliff could collapse next

Drone surveys have helped University of Tartu geologists identify dangerous sections of the Pakri cliff before they collapse, ERR reports. The 3D models show how seawater erodes softer sandstone beneath layers of limestone, eventually causing the rock above to fall into the sea.

Two winters ago, in mid-January, a viewing platform fell from the cliff into the sea. A 30-metre section of the edge broke away, and the cliff retreated 12 metres in one collapse. About 2,800 cubic metres of rock fell into the water, equivalent to roughly 160 dump trucks. Although the collapse appeared sudden from above, it may have been possible to anticipate from the side, writes Tiia-Linda Purge, a junior researcher in geology at the University of Tartu.

Near Paldiski, the Baltic Klint rises directly from the sea to a height of up to 25 metres. An old lighthouse dating from the time of Peter I stands on the edge, alongside paths worn by visitors. Geologists have studied the cliff’s rocks and layers for more than two centuries, but its retreat has mostly been measured from above by comparing old maps and aerial photographs.

The cliff’s upper seven metres consist of limestone. The rock is hard and fractured at the top, while its lower layers are more clay-rich, sandy and crumbly. Beneath it lie green glauconite sandstone, studied as a raw material for potassium fertiliser, black shale known as graptolite argillite, and softer sandstone at the bottom. These rocks formed on the floor of an ancient sea nearly half a billion years ago.

To study the parts of the cliff face that cannot be seen from above, geologists took drone photographs in 2015, 2022 and 2024. Software combined the images into 3D models that allow researchers to examine and measure rock layers, cracks and caves. They also mapped the cliff edge using six aerial photographs covering the years from 1959 to 2024.

Taking the photographs was challenging: the drone had to fly just metres from the cliff face and the water while avoiding fallen rock. Waves and debris prevented the team from photographing the base as closely as they wanted, and the ceilings of deeper caves remained out of view. The researchers used photogrammetry, in which computer algorithms match points in photographs taken from different angles to create a 3D model.