Mass strandings, in which groups of otherwise healthy whales swim ashore and die, have puzzled observers for centuries. Ancient writers attributed the events to divine displeasure; nineteenth-century naturalists preferred to blame poor eyesight. Modern marine biologists have a richer, though still incomplete, set of explanations. Most concern the ways in which whales use sound to navigate, and the ways in which human activity, submarine topography and social behaviour can disrupt that navigation.

The species most frequently involved in mass strandings are toothed whales, especially beaked whales, pilot whales and, less often, sperm whales. All of these animals rely heavily on echolocation, producing clicks and listening to the returning echoes to build up a picture of their surroundings. Baleen whales such as the blue whale or the humpback, which use sound mainly for communication rather than fine-scale navigation, rarely strand in large groups. This simple pattern suggests that whatever causes mass strandings is connected to the way toothed whales perceive their world.

One widely discussed factor is sloping, sandy coastline. Echolocation clicks travel well through deep water and bounce clearly off hard surfaces, but a gently shelving beach covered in soft sand returns a weak and ambiguous echo. A whale approaching such a shore may not realise how close the bottom is until it is already aground. Maps of strandings in New Zealand, compiled by Dr. Ana Marquez at the University of Auckland, show a strong concentration at exactly this kind of coastline, particularly at Farewell Spit and parts of the Chatham Islands. Places with steep, rocky shores, where echoes are sharp, record far fewer events.

Topography alone, however, cannot explain why healthy animals continue into shallow water when they could turn back. Here the social behaviour of the species becomes important. Pilot whales live in tight-knit groups, and tagging studies have shown that individuals remain close to their relatives for most of their lives. When one animal becomes disoriented, others apparently follow. Rescuers who refloat a single whale often find that it swims directly back towards the rest of the group on the beach, a behaviour that is difficult to reconcile with self-preservation but consistent with strong family bonds. Dr. Marquez describes these animals as 'socially anchored' to their companions.

Since the 1990s, a third factor has entered the debate: human-made noise. Military sonar, used to detect submarines, emits very loud pulses in the same frequency range as the hearing of many beaked whales. A series of strandings in the Bahamas in 2000, coincident with naval exercises, provided the clearest evidence. Autopsies revealed gas bubbles in the tissues of the stranded animals, a pattern consistent with the decompression illness familiar to human divers. The current hypothesis is that the sonar pulses startled the whales into rising too quickly from deep dives. Similar events have since been recorded in the Canary Islands and off the coast of Greece, and temporary restrictions on sonar use in known whale habitats are now standard practice in several navies.

It is tempting to blame strandings on a single cause, but the evidence increasingly points to combinations. A beaked whale that has been exposed to sonar may surface in an unfamiliar area, lose contact with its group, and drift towards a sandy shore whose echoes it cannot interpret. None of these factors alone would necessarily kill it; together they may be fatal. This interaction makes prediction very difficult, and it also means that reducing one cause, such as sonar use, may have a larger effect than the size of that cause alone would suggest.

Rescuing stranded whales is correspondingly challenging. Volunteer teams in Tasmania and Cape Cod have become highly organised, using stretchers and rafts to return individuals to deeper water. Success rates are modest. Even when animals are physically healthy when refloated, they often restrand within hours. Some rescue groups now attempt to move an entire pod together, on the assumption that social bonds will hold them on a chosen course. Early trials by the Tasmanian Whale Response Network suggest that this group-release method roughly doubles the proportion of animals that remain at sea, though the sample is still small.

Strandings will probably never be eliminated. They appear to be part of the natural history of toothed whales, exaggerated in some places and periods by human noise and by the shape of particular coastlines. But a clearer picture of why they happen has already changed how they are managed. Instead of treating each event as an isolated tragedy, researchers now record it as data, feeding it into an evolving model that may one day allow at-risk groups to be identified before they ever reach the beach.