Warning compliance

[nytimes.com]

Authorities on the densely populated Indonesian island of Java concluded in mid-October that the threat was imminent enough to require sending troops to forcibly evacuate tens of thousands of villagers living on the mountain’s slopes, directly in the way of volcanic ash falls, mudslides and perhaps even lava flows …

that didn’t come. The government said on Thursday that the threat had now subsided enough for most evacuees to return to their homes and lands, and learn whether they had been looted or ruined over the weeks they were left untended.

In just two days, we’ve gotten two more big datapoints for the age-old quandary facing public officials around the world about where to set the threshold for public warnings of less-than-certain disaster.

There doesn’t seem to have been a crying-wolf issue in either case: both Mount Kelud eruptions and North Sea storm-surge floods have wrought devastation in living memory, and the authorities could offer plenty of objective physical grounds for their concerns.

Still, erring on the safe side takes its own toll, both material — the evacuated Indonesians apparently had ample cause to worry about looting — and psychological. Even when they are issued in good faith for good reason, every false alarm can drain some of the menace, and some of the effectiveness, out of the next warning.