A group of American scientists are developing a lightning forecast system that could predict which storm clouds are likely to produce lightning and also indicate its duration.

While there is no operational lightning forecast system using radar, researchers using the existing Doppler weather radar system can get lightning predictions right about 90 percent of the time, but can only give about a 10 to 15 minute lead time.

Supported by a two-year research grant from NASA, scientists in the Earth System Science Center at The University of Alabama in Huntsville (UAH) are combining data from weather satellites with Doppler radar and numerical models in a system that might warn which specific pop-up storm clouds are likely to produce lightning and when that lightning is likely to begin and end, reported Science Daily.

It has been estimated that in an average year lightning kills about 24,000 people while injuring another 240,000 worlwide.

"One of our major goals is to increase the lead time that forecasters have for predicting which clouds are most likely to produce lightning and when lightning will start," said John Mecikalski, one of the project directors and an associate professor in UAH's Atmospheric Science Department.

"If we can combine data from satellites, radar and models into a single lightning forecast system, we can give the National Weather Service and other meteorologists a new tool to support forecasts."

In addition to using cloud top temperature data available through existing weather satellites, the new lightning forecast system is also being designed to use lightning flash information gathered by the Geostationary Lightning Mapper, an optical instrument slated to be launched aboard the next generation of NOAA weather satellites in 2016.