Can Lightning Predict Hurricanes?

By MICHAEL KANELLOS Radar and satellites can track the path of storms, but Earth Networks, a company that operates networks of environmental sensors, says it can predict their intensity. Sensors detect the location of electromagnetic activity and send the data to supercomputers….

By MICHAEL KANELLOS

Radar and satellites can track the path of storms, but Earth Networks, a company that operates networks of environmental sensors, says it can predict their intensity.

Sensors detect the location of electromagnetic activity and send the data to supercomputers.

The company has come up with a system that detects and geographically pinpoints electrical activity, that is, lightning, in storms. It then runs this information along with other meteorological data on temperature, pressure and wind speed through supercomputers to predict whether a storm will peter out as it hits land or wreak even greater havoc.

“We measure every spark in the cloud,” says the company’s chief executive, Bob Marshall. “Radar is great at forecasting the track of a storm, but intensity is more of a mystery. Predicting intensity is not much better than it was a decade or two ago.”

The company has already installed about 360lightning sensors. Most are in the United States and function as part of its 8,000-node domestic weather network. Soon, Earth Networks plans to set up a lightning network in Brazil, where storm tracking is already hampered by spotty satellite and radar coverage. The company has also had discussions with Indian aviation officials.

There’s no denying the weather has become more extreme. Years of drought in Australia have been followed by devastating floods. Weather events causing more than $10 billion in damage, from floods and wildfires to an East Coast snopocalypse, have hit the United States this year. In Pakistan, large swaths of land sit under floodwaters.

Munich Re, the large reinsurance company, has estimated that natural disasters caused $265 billion in damage in the first six months of the year, five times more than the average losses for that period over the last decade.

Investors and entrepreneurs have begun to inch into the market for weather technology. WeatherBill, an insurer that has received $42 million from sources like the venture capital firm NEAGoogle Ventures and others, harnesses data analysis to underwrite climate insurance policies for farmers.Liquid Robotics is meanwhile creating a flotilla of robotic ships that will monitor the high seas.

Earth Networks started in the early 1990s with WeatherBug, a PC-based weather app. In 2006 it began to collaborate with M.I.T. on lightning detection and introduced the first network for that purpose in 2009.

How does the system work? Lightning produces a brief electromagnetic pulsethat is detected by the company’s lightning sensors, essentially large antennas connected to computers. By analyzing the times that various sensors detect pulses, Earth Networks can triangulate the location and path of electrical activity. Think of it as a speed trap for lightning.

A concentration of electrical activity at the eye of a storm is often a harbinger of destruction. If lightning moves from the eye to the periphery of the cloud, the energy is probably dissipating.

By studying lightning activity in Hurricane Irene, the company predicted on Aug. 24, when the Atlantic Seaboard was still on high alert, that it would not be as severe as initially anticipated because the electrical activity had already moved from the center to the edge of the storm.

And the company asserts that its lightning networks could have given residents of Springfield, Mass., a 50-minute warning in advance of a tornado that struck on June 1. Instead, residents were only warned a few minutes ahead of time.

A network node, a mini-weather station equipped with meteorological sensors, costs about $10,000, Mr. Marshall said. Earth Networks owns the networks and sells data services to subscribers. In addition to some government agencies, utilities subscribe to better predict power losses and manage repair crews.

Ultimately, Earth Networks wants to devise other sensors and new ways of mining weather data to possibly provide longer-range forecasts — flood warnings 30 days in advance, for example, or predictions on the severity of heat waves. It is in the midst of building a network of sensors for carbon dioxide flows that could also help measure the effectiveness of forests in absorbing greenhouse gases.

Lightning science remains a relatively new field. Mr. Marshall says that early lightning data can’t precisely predict what will occur but that enough of a correlation exists to give emergency services groups and government agencies some advance warning.

“We want to put the earth through a battery of tests,” he said.