The Tōhoku Domino Disaster

Advanced computing is essential to disaster response, and the ongoing difficulties in Japan are no exception.

For some background on the events in Japan, check out our interactive timeline below. Then read on to learn about some of the computing efforts that are helping to make a difference.

When the Earth shook, so did at least one of over a thousand seismometers scattered across Japan by the Japan Meteorological Agency and the National Research Institute for Earth Science and Disaster Prevention (NIED). They sent their data to the JMA, where computers analyzed the data and sent out a warning that would be broadcast via cellphones, televisions, websites, and radios: prepare yourself.

The first domino in a long series of disastrous events was tipped.

Thousands died, but the toll could have been much more if it hadn't been for the JMA early warning system.

At the same time, the US National Oceanic and Atmospheric Administration's Center for Tsunami Research at the Pacific Marine Environmental Laboratory (PMEL) took the seismic data from the earthquake and plugged it into the Method of Splitting Tsunami (MOST) model to predict the ensuing tsunami. (A similar model called TUNAMI N2 was developed by researchers at Tohoku University; it is still in use in a variety of contexts.)

When the ocean moved, so did the Deep-ocean Assessment and Reporting of Tsunamis (DART) buoys deployed by NOAA. As the tsunami reaches each of these high-tech buoys, they record water levels and send their data back to PMEL, where the technology was first developed. There, the MOST model uses the buoys' data to refine its predictions for the tsunami. (You can see graphs comparing the predictions with the data in our timeline, above.)

Then, the ocean rose up and subsumed the sea wall near Fukushima I Nuclear Power Station (also known as Fukushima Daiichi). The earthquake had already knocked out the electrical lines that were powering the cooling system. The tsunami flooded the backup diesel generators, disabling them indefinitely.

Engineers have struggled since then to keep the cooling system running, with limited success. Even though the three running reactors located at Fukushima Daiichi shut down automatically when the earthquake was detected, both the cooling fuel rods and the spent fuel rods are in danger of causing a meltdown if they are not continually cooled.

The result is that increasing levels of radiation have escaped.

Here, too, computing can serve an invaluable purpose.

On 15 March, a team of 33 experts from the US National Nuclear Security Administration joined six Department of Energy experts already in Japan.

Since arriving, the team has been assessing, surveying, monitoring, and sampling for radiation using over 17,200 pounds of equipment they brought along with them.

That equipment includes the Aerial Monitoring System, which the team has used to gather data during over 40 hours of flights. The AMS data has been combined with data from over a thousand ground  monitoring points; the Department of Energy's analysis can be seen in the slideshow above, or you can download it as a Powerpoint presentation here.

"That data, along with additional information gathered from other agencies and our partners in Japan, is sent to NARAC," said NNSA representative Damien LaVera. "There, it is combined with data about atmospheric and weather conditions to develop models that can be used to predict the outcome of various scenarios."

According to LaVera, NARAC (the National Atmospheric Release Advisory Center at Lawrence Livermore National Laboratory) uses the NNSA supercomputers at LLNL. However, an LLNL representative told iSGTW that LLNL supercomputers are not being used in the disaster response. At press time, these conflicting reports had not been reconciled.

The NARAC team, which specializes in disaster response in the case of airborne hazards, feed the data they receive into their existing models to "predict the outcome of various scenarios," LaVera said. "As an example, NARAC can take data gathered on the ground or from other sources and predict the way that changes in the weather could impact the distribution of contaminants."

In this case, the contaminant they are modeling is radiation.

The teams on the ground in Japan and NARAC were put in play to provide the US government with predictive modeling capability to aid in decision making. The data gathered and predictive model results are also being shared with Japanese partner organizations. By working together, scientists, engineers, and government agency representatives from around the world will hopefully be able to minimize any further damage. The cascade of dominoes stops here.