Satellites get the jump on storm prediction

When it comes to severe thunderstorms, every minute of advance warning can be vital. Present methods rely on radar to detect impending storms, but a new technique that uses satellites to measure the temperature changes in the tops of clouds, could predict severe thunderstorms up to 45 minutes earlier than relying on traditional radar alone.

Clouds start cooling long before radar can identify them as storms. As a warm cumulus cloud grows and expands upward into higher altitudes, it will rapidly cool. Rapid cloud-top cooling indicates that a cloud top is rising into the frigid upper reaches of the atmosphere and can reveal the formation of a severe storm.

Cloud temperatures can be measured by the wavelengths of light they radiate in the near-infrared and infrared frequencies. Current geostationary satellites over the U.S. can discern five different bands in these frequencies, each band revealing a different state of cloud development. Looking down from space, the satellite can determine whether the cloud top consists of liquid water, supercooled water or even ice.

Scientists from the Cooperative Institute for Meteorological Satellite Studies (CIMSS) found that by running high-speed five-minute satellite scans through a specially designed computer algorithm, they were able to quickly analyze cloud top temperature changes to look for signs of storm formation.

"The value of detecting and analyzing these changes is that we can get up to a 45-minute jump on radar detection of the same storm system. A 'nowcast' becomes a 'forecast,'" says CIMSS scientist Wayne Feltz.

Feltz and other CIMSS colleagues demonstrated their "Convective Initiation Nowcast" and "Cloud Top Cooling Rate" products at NOAA's annual Hazardous Weather Testbed (HWT).