(Volcano Watch is a weekly article written by scientists at the U.S. Geological Survey’s Hawaiian Volcano Observatory.)

Footsteps. Jumping up and down. Wind, machinery, traffic, surf. Many things shake the ground. Recording on seismographic equipment the seismic waves produced by explosions is a well-known technique to explore the Earth for oil. Recording earthquakes is central to seismic monitoring for understanding earthquakes, forecasting tsunamis, and monitoring active volcanoes.

Success requires that these desired signals be distinguished from the noise signals from other sources, like wind and surf, but improved technical tools are opening up new monitoring possibilities.

Since the study of earthquakes has matured both observationally and theoretically, the use of computers and digital technology have sped up the analysis and fundamentally transformed seismology.

At one time, it was an immense computer challenge to simply acquire and handle large amounts of earthquake data from moderate to large seismic networks like those operated by the U.S. Geological Survey in California and at Hawaiian Volcano Observatory (HVO).

Brought online in 1986 to more fully automate the processing steps, the first computer systems to directly collect and analyze seismic data at HVO boasted 8 megabytes (Mb) of memory and 456 Mb of storage disk space.

The computer software that ran on this system was conceived, designed, and written by a small team in southern California led by Carl Johnson, who, at the time, was a USGS scientist. Dr. Johnson is widely regarded as a visionary and genius in computers and scientific computing.

The HVO systems, and similar systems used by the USGS in California, though small and limited by today’s standards, were then state-of-the-art. Large seismic networks like those here or in California generate several gigabytes of digital seismograms each day.

To achieve the earthquake monitoring mission most efficiently and to steer clear of quickly filling up storage disks (expensive and limited in capacity compared to today’s disks), software was designed to recognize and capture earthquakes from the continuous digital streams obtained from the seismic instruments.

Once the earthquakes — or event triggers — were captured, digital analysis focused on these signals, and extensive research has resulted in important advances in earthquake science. At the same time, most data – collected between the earthquakes of interest -simply went onto tapes or other suitable storage media and were largely unused.

Through time, the evolution of computers and their capabilities has made it feasible to handle massive amounts of data, and to repeatedly and automatically process it. Over the last decade, researchers have demonstrated that ambient seismic noise — recorded between earthquakes and other events of interest — also carries important information about the Earth.

Recently, scientists studying Le Piton de la Fournaise Volcano on Reunion Island in the Indian Ocean have made important progress using these ambient signals. Cross-correlating ambient seismic noise between seismographic stations through time has allowed them to look more closely at processes within the volcano, as well as times when the volcano is not seismically active.

The basic products of ambient seismic noise cross-correlation are estimates of relative changes in seismic wave speeds between pairs of seismic stations. With amazing precision to fractions of a percent, the scientists are able to track changes in the volcano through time.

Before recent eruptions at Le Piton de la Fournaise, from 1999 to 2000 and from 2006 to 2007, clear decreases in seismic wave speeds were observed over periods ranging from weeks to days.

This is particularly impressive, since the seismic crises associated with these eruptions are significantly shorter in duration, ranging from half an hour to two days. The changes in wave speeds are thought to be related to stresses and cracking due to magma movement within the volcano.

Additional research and analysis are required to determine the locations of where the velocity changes are occurring, and work is currently underway to render these techniques into real-time volcano monitoring and forecasting tools at Le Piton de la Fournaise.

These are important and exciting results, and we anxiously look forward to future findings of our colleagues. Ambient seismic noise techniques that take advantage of a much larger amount of our recorded data will enable us to more consistently track and map changes occurring within the volcanoes.

Active volcanoes are as dynamic and constantly evolving as the techniques used to monitor them.