April 20, 1999 Tsunami: Kasu Cone, Papua New Guinea

Reference #14091:

[…] a 150,000-m3 landslide into its lake. Displacing 5–10% of the lake’s volume, the landslide induced a 15-m-high wave which flattened vegetation throughout the crater interior, and broke over a low saddle in the crater’s east rim where it destroyed two houses and injured 11 people, killing one.

The landslide was probably triggered by heavy rains, with permeation of the slope aided by desiccation cracks from the preceding year’s drought.

Two thatch houses in the saddle, just 3 m above lake level, were washed away. Injuries to the occupants were described as broken bones.

The height of vegetation disturbance ranges from 15 m above the lake in most areas, down to lake level immediately adjacent to the landslide. The flattened kunai points away from the landslide along the eastern and western crater walls, but points down towards the lake on the northern wall.

Villagers living on the cone’s north rim and flank said that on 20 April 1999 they were awoken at 03:00 hours by a loud noise described contradictorily as both a low humming and a loud thundering. They then discovered that the two houses in the saddle had been destroyed by water from the lake, and the occupants washed down the eastern flank. They also noticed an “unpleasant” odor in the air, and described the lake as being empty. Two hours later at 05:00 hours, some described feeling groundshaking and hearing some noise. Still an hour before dawn, they could not yet see the crater’s south end. After sunrise, they observed the landslide scar and most attributed it to the 05:00 ground shaking. By 10:00 hours they state that water had “bubbled” back into the lake.

The villagers, however, insisted that the landslide occurred two hours after the houses were destroyed. Nonetheless, other mechanisms for water discharge seem implausible. Renewed magmatic activity is not supported by any of our observations, no recent eruptives, no ongoing seismic activity, and moderate pH and temperature of lake waters. Gas eruption, which generated a 100-m-high water fountain at Lake Nyos (Kling et al. 1987), seems similarly unlikely, since Kasu’s shallow depth precludes storage of sufficient gas to drive eruption. Seiche is also excluded by the lack of earthquakes on the night of the event, and the pattern of vegetation disturbance. The flattened Kunai points away from the landslide on the adjacent crater walls, indicating a directed source and inconsistent with the oscillations of a seiche or the other mechanisms.

We conclude that the events at Kasu Cone were initiated by a landslide from the crater’s southern wall at 03:00 hours.