Disaster Forensics

It’s my job to drive straight into the heart of disaster zones.

On 11 March, a 9.0-magnitude earthquake triggered a monstrous tsunami that smashed into Japan’s northeast coast, killing more than 15 000 people in minutes and reducing entire towns to rubble. In the days that followed, more than 80 000 Japanese citizens fled their homes after the tsunami started a meltdown at three of the reactors at the Fukushima Dai-ichi nuclear power station. Those citizens left their whole lives behind, and most are still living as refugees. But in early April, I drove into the wreckage of Japan’s coastal towns to see what lessons I could learn in the ruins.

As an electrical engineer with a keen interest in what I call ”disaster forensics,” I travel to the worst natural disaster sites around the world to assess the damage inflicted on communication networks and electric power grids. I’ve surveyed the aftermath of three major Gulf Coast hurricanes, including Katrina, and I’ve stood in the rubble caused by earthquakes in Chile, New Zealand, and Japan. As I’ve collected field data, I’ve begun to challenge the common belief that humans can’t compete with nature’s fury and that most of our creations will fail in a hurricane’s winds or a tsunami’s waves. That fatalism doesn’t sit well with me. I think that studying the world’s worst natural disasters can lead to better designs and critical infrastructures that can better withstand the brunt of a storm or the upheaval of an earthquake.

Of all the assessments I’ve performed, my April 2011 trip to Japan was undoubtedly the most challenging—and the most heart wrenching. At that time, the damaged nuclear reactors at Fukushima Dai-ichi had not been fully stabilized, and the extent of the area contaminated by radiation fallout was not clear. I drove a circuitous route on the west side of Japan’s main island because the Sendai Airport had just started limited operations, the ”bullet” train wasn’t yet back in service, and the highway from Tokyo ran within the U.S. State Department’s recommended evacuation zone. I carried all the food I would need for my five days on the road and water for my entire 10-day trip, because I’d heard stories of contaminated tap water and shortages of bottled water. I also carried a dosimeter to measure radiation levels everywhere I went and quickly left areas that set off the dosimeter.

Photo: Alexis Kwasinski

The 11 March tsunami in Japan destroyed a base station and cell tower near Rikuzentakata.

When I started my assessment, I was surprised to find that ground-shaking damage was relatively minor for a 9.0-magnitude earthquake. But the tsunami damage was shocking. As I stood next to a damaged cell tower on a hill about 20 meters above sea level near the town of Ryoishi, I could only imagine the horror of the people who ran from the tsunami and found no safety even on that high hill. In the towns of Otsuchi, Rikuzentakata, and Onagawa I walked through eerie silences and gazed at the piles of debris that towered over my head. In Otsuchi, I stood in front of a destroyed firehouse and thought of the firefighters who died after closing the seawall gates in a desperate attempt to stop the onrushing water.

Photo: Alexis Kwasinski

The telecommunications central office in Rikuzentakata was one of the few buildings still standing after the tsunami.

Amid this destruction, I looked carefully at the structures that survived. I found that the central office buildings belonging to Japan’s biggest telecom company, NTT, were often still standing, although they had been submerged by the tsunami and had been hit by floating debris. In Onagawa the remains of a home ended up perched on top of a two-story NTT building. These buildings housed the switches that routed telecom data, and in Japan they were designed to withstand the most powerful earthquakes. The buildings also had some precautions against tsunamis—such as watertight seals around the doors on the ground floor—but these safety measures were insufficient protection, because no one anticipated a tsunami of such extraordinary height. On 11 March, water poured into the windows on the upper floors of the buildings and soaked the equipment inside.

Photo: Alexis Kwasinski

After the tsunami, NTT’s central office building in Onagawa had the remains of a house on its roof.

In Japan, I saw the inherent fragility of the power grid. There were extensive outages not only in the heavily tsunami-damaged coastal areas but also in the lightly affected inland regions. I also came up against some frustrating facts of life regarding residential solar power. Japan has a relatively high number of homes equipped with photovoltaic (PV) solar panels, and many inland homes equipped with these PV panels could theoretically have remained powered during the day. Unfortunately, residential PV-grid interconnection standards prevent PV panels from providing power during a grid outage [PDF], so these homes got no benefit from the panels on their own rooftops.

This may seem like a wasted opportunity—and maybe it is—but utilities say that homeowners shouldn’t be allowed to disconnect from the grid and power their households with solar panels during an outage due to safety issues. Utilities worry that if the disconnection from the grid is incomplete, a lineman could go out to work on damaged power lines thinking they’re safe when in fact they have power running through them from a solar panel somewhere. This argument is controversial, because technical solutions exist to ensure a proper disconnection and thus solve any safety problems. Unfortunately, these solutions require technology that isn’t in common use right now and therefore might incur higher costs. Almost all conventional ”off-the-shelf” inverter controllers, which are used to manage the electricity from residential solar panels, prevent their operation unless the main power grid is present and live.

Photo: Alexis Kwasinski

This religious center in Kamaishi, which was used as an emergency shelter after the tsunami, couldn’t draw power from its solar panels or wind generator, which were left undamaged.

Nevertheless, there are some ways to ”go local” with energy production under the current rules. In the Japanese city of Sendai, a small local power grid, called a microgrid, showed its potential during the blackout. Powered by natural-gas engine generators and solar panels, the microgrid was able to keep the lights on at Tohoku Fukushi University and a neighboring hospital throughout the blackout. While under normal circumstances the Sendai microgrid is connected to the city’s macrogrid, its ability to isolate itself and keep generating power is valuable to customers who need a highly reliable source of electricity.

I got my start in disaster forensics in October 2005, when I took part in a National Science Foundation study on the damage caused by Hurricane Katrina. That enormous and devastating hurricane barreled into the Gulf Coast in late August of 2005. The furious winds and extensive flooding caused damage amounting to US $108 billion, more than any other hurricane in U.S. history. I had the assignment of collecting field data to determine why and how the storm had wreaked such havoc on critical communication and electric power infrastructures.

I drove south from Champaign, Ill., where I was completing my Ph.D. in electrical engineering, and headed for the most heavily damaged areas of southern Louisiana. On the east side of the Mississippi River Delta, I turned onto a narrow highway that runs through endless marshlands and found my way to the tiny towns of Yscloskey, Delacroix, and Point à la Hache. The debris-filled streets were silent—the people who lived there, mostly fishermen and workers in the offshore oil industry, hadn’t been able to return. Katrina had pushed huge swells of ocean water up coastal rivers and canals, and that storm surge submerged this low-lying delta area and destroyed most of the homes. Smashed cars and boats poked out of the canals, mixed with uprooted trees and clumps of seaweed.

Photo: Alexis Kwasinski

Hurricane Katrina didn’t leave much of Delacroix, La., intact.

When I began studying the power grid, I was surprised to find that very light damage—in some places, less than 1 percent of components—had still resulted in total blackouts for large geographic areas. For example, in some places a single toppled utility pole or a broken switch at a substation had brought down a wide swath of the network.

Problems in the grid rippled outward to the telecom system: Katrina left many cellular towers intact, but the power problems quickly took a toll. During the blackout, intact cell towers switched to backup battery and diesel generator power. But within a day or two, the cell towers began blinking out of service: Their batteries were depleted, their generators were out of fuel, and workers couldn’t travel through the disaster zone to reach them. When workers finally did arrive at the failed cell towers, they deployed multiple portable generators at each site to power the telecom base stations at the towers. However, that only created a more complicated problem: Workers then faced the daunting logistics of refueling hundreds or even thousands of generators for several weeks. One solution I can see to this problem in some sites would be to rely on alternative local sources of energy, such as solar panels or small wind generators, in order to reduce the load on the portable generators. Small permanent generators connected to city gas lines could provide an even better solution for these sites.

Photo: Alexis Kwasinski

Multiple portable generators powered various base stations sharing a cell tower in Louisiana after Katrina. Each generator needed to be refueled about every 24 to 48 hours.

Katrina also destroyed the telecom central office buildings housing equipment that routed landline and wireless communications through nearby towns. As telecom companies tried to restore service, they faced a grim question unique to the world’s worst disaster zones: How much service needed to be restored? Most of the demand for telecom service in the storm-tossed areas of the Gulf Coast had disappeared because the residents had fled their ravaged homes—or they had no homes to return to. So BellSouth, the primary operating company in southern Louisiana, decided not to rebuild many central offices in damaged areas and instead deployed small digital loop carrier (DLC) system cabinets. These cabinets transmit phone calls and Internet data to distant central offices through fiber-optic cables, which have greater capacity than conventional copper lines. When placed on high platforms to avoid future floods, these cabinets can form part of a resilient telecom system.

Photo: Alexis Kwasinski

The central office building in Yscloskey, La., was destroyed by Katrina and wasn’t rebuilt.

Photo: Alexis Kwasinski

A DLC cabinet system [left], seen here after Hurricane Gustav in 2008, replaced the Yscloskey central office building. The seaweed draping the fence was carried there by the hurricane’s storm surge.

However, as I saw three years later, when I surveyed the wreckage that Hurricane Ike caused in Texas, DLC cabinets have problems of their own.

Hurricane Ike, which struck the Gulf shores in September 2008, hit with great fury and caused a total of $30 billion of damage in Texas, Louisiana, and beyond, making it the third-costliest hurricane in U.S. history (after only Katrina and 1992’s Andrew). I’ll never forget the coastal town of Gilchrist, Texas, where one lonely home remained standing. Power outages extended from the battered coastline more than 100 miles inland and caused widespread telecom outages (the storm’s remnants went on to cause problems all the way up to Canada). In Texas, DLC cabinets that routed landline communication data were largely responsible for the telecom problems.

DLC cabinets have the same problem as cell towers—in a blackout, their batteries give out within a few hours. (Telecom central offices typically have permanent generators that can keep power flowing for at least 72 hours, which is a reasonable cushion.) In order to restore service, power companies distributed hundreds of portable generators to DLC cabinets throughout east Texas; this gesture created the extra logistical burden of keeping them fueled. Since the spread of broadband communication systems is increasing the use of similar curbside cabinets, it’s a safe bet that this problem will become more common in future disasters.

Photo: Alexis Kwasinski

Gilchrist, Texas, was knocked flat by Hurricane Ike.

Downtown Houston power outages were not as severe as in neighboring areas, thanks to the partially buried power grid. Buried power lines are expensive enough to limit their use, but they could be a solution for relatively small but particularly vulnerable areas, such as the Bolivar Peninsula and Galveston Island in Texas. Burying the single transmission line that serves the peninsula would have prevented multiple failures due to broken and fallen poles. Doing so would also have reduced blackouts during the critical evacuation period—the line failed several hours before the hurricane made landfall—when power was needed for traffic lights and gas stations.

I was also struck by the scattershot, uneven nature of the severe damage. Just a few miles from Gilchrist, the town of High Island made it through the storm with only light damage. This fact suggests that engineers face an important challenge: to build an overall infrastructure that is resilient and flexible, so that when serious problems arise in one small area, the rest of the network doesn’t go down.

Photo: Alexis Kwasinski

Poles snapped by Ike brought down the only power line serving the Bolivar Peninsula.

As I’ve worked in this field of disaster forensics, I’ve come to think of disasters as undesirable stress experiments that allow us to evaluate the performance of critical infrastructures. The lessons we learn may allow us to save lives and reduce the economic impact of future disasters. A category 5 hurricane is statistically overdue in Florida, and major earthquakes have long been expected near Tokyo and on the United States’ Pacific Northwest coast, where a tsunami is also likely.

It’s not just communications and power engineers who can learn from my recent tours of the world’s worst catastrophes. For example, the successful construction practices observed in NTT’s central office buildings in Japan could be replicated along the Pacific Northwest coast to reinforce existing emergency facilities and school classrooms. Japan’s tsunami evacuation structures for coastal communities with distant high-ground areas could be duplicated in the United States. Perhaps the main lesson that I took from my travels is that it’s simply not true that little can be done to prevent disasters. But it’s up to the communities at risk, from government officials to individual citizens, to demand the investments that will prepare their towns and cities for the worst. If these communities don’t learn from those who have suffered before, history will tragically repeat itself.

Photo: Alexis Kwasinski

The tsunami evacuation structure seen in the background saved lives in Minami Sanriku, Japan. Such buildings could also save lives along the U.S. Pacific Northwest coast.