Skip to main content
U.S. flag

An official website of the United States government

Volcano Watch — How much warning should we expect before big eruptions?

September 9, 2010

How long is the warning between the first rumblings of a volcano and its eruption? This haunting question was raised anew with the eruption of the Indonesian volcano Mount Sinabung on August 29, 2010.

Sinabung is a small, composite, andesitic volcano in northern Sumatra. Its summit, 2,460 m (8,071 ft), rises only a thousand meters or so above the surrounding terrain, much of which is dotted with numerous small farms.

Initial reports suggested that Sinabung lay in repose for 400 years or more. An explosion may have occurred in 1881, and fumarolic (steam vent) activity persisted until 1912. But were there more recent events that might have hinted at Sinabung's unrest?

Swarms of earthquakes and tremor, followed by explosions--each progressively larger--were detected near the summit on August 29 and 31 and on September 3 and 7. The earthquakes were monitored, warnings issued, and citizens evacuated. The September eruptions were forecast by Indonesia's Center for Volcanic and Geologic Hazard Mitigation. They appear to have been mostly phreatic, that is, resulting from the interaction of ground water and hot rock.

Some news reports mentioned the glow of incandescence, visible at night, when magma reached the surface or when hot gases heated the explosions' rock debris to 450 degrees C or more. Plumes of ash during the largest explosive event on September 7 rose 5,000 m (16,000 ft or 3 mi) high. By then, as many as 20,000-30,000 people had been evacuated. Other warning signs may have preceded the September explosions; but interpreting them in what was a poorly monitored area is fraught with difficulty.

We can look at the historical record for premonitory periods of other composite volcanoes. Mount St. Helens, which had previously erupted in A.D. 1857, underwent two months of seismic unrest, extensive surface deformation, and a series of phreatic explosions preceding its collapse and plinian eruption on May 18, 1980.

The Philippines' Mount Pinatubo had a lengthy repose period--about 450 years (like Sinabung's). It sustained about two months of unrest, gas emissions, and phreatic explosions (like Mount St. Helens) before its June 1991 eruption, the second largest of the 20th century.

In May 2008, Chiaten volcano in Chile awoke from about 9,600 years of slumber--with only a few days of seismic warning and felt earthquakes--followed by one of the largest eruptions of the past two decades. As with Sinabung, Chiaten lacked a close-in network of seismometers to detect microearthquakes, quakes too small to be detected except by nearby sensitive instruments--and typically the first signs of a volcano's unrest. Volcanologists suspect that a close-in network at Chaitén might have provided weeks or months of warning.

Notably short warning marked the 1989 eruption of Mount Redoubt, Alaska, which proceeded from background seismicity to full eruption in 24 hours. Redoubt's previous eruption occurred only 21 years earlier, ending in 1968. Compared to the centuries of rest that preceded the reawakening of Pinatubo and Sinabung, its repose period was brief.

Volcanologists are aware of the potential for "open-system" (frequently active) volcanoes like Redoubt to progress quickly from unrest to eruption. In contrast, "closed-system" volcanoes (inactive for many decades or centuries, like Mount St. Helens, Pinatubo, and Sinabung) are expected to give substantial warning before eruption. At closed-system volcanoes, magma must establish a new pathway through the upper crust--a process that fractures rock, generates earthquakes, and deforms the volcanic edifice.

Perhaps Sinabung's eruption is a premonitory phase, to be followed in a few months by more substantial eruptive activity. Or maybe the explosions of September 2010 will close this episode without full-blown eruption, reminding us that this volcano, capable of renewed activity after long repose, is worthy of more detailed monitoring.


Volcano Activity Update

Very small breakouts have been sporadically active at the base of the pali over the past week, but otherwise, no significant breakouts have been observed on the flow field. On Saturday, September 4, a small lava pond formed in the west portion of Pu‘u ‘Ō‘ō crater, and this pond remains active at the time of writing (Thursday, September 9).

At Kīlauea's summit, a circulating lava pond deep in the collapse pit within the floor of Halema‘uma‘u Crater has also been visible via the Webcam throughout the past week. The lava level has slowly risen with the ongoing inflation, and this slow change has occasionally been interrupted by abrupt increases in the height of the lava surface. These periods of high lava level were short-lived, lasting up to several hours, and ended with a sudden drop of the lava surface, back to its previous level. Volcanic gas emissions remain elevated, resulting in high concentrations of sulfur dioxide downwind.

One earthquake beneath Hawai‘i Island was reported felt during the past week. A magnitude-2.5 earthquake occurred at 4:20 a.m. on Wednesday, September 8, 2010, H.s.t., and was located 5 km (3 miles) southeast of Kīlauea summit at a depth of 4 km (2 miles).