In-flight turbulence is hard to predict and the leading cause of injuries on airplanes. Researchers at the National Center for Atmospheric Research in Boulder, CO, have developed software to improve turbulence prediction. When flying through storms, pilots use onboard Doppler radars to scout for drier areas, assuming that these will be calmer. But turbulence can still strike. Algorithms in the new software reduce noise in the radar data; warning algorithms then analyze the data to find the amount of movement in the tiny bits of water and ice found even in “dry” areas. The wider the spectrum of velocities among the droplets, the more likely an encounter with turbulence. In flights on a NASA test plane, the software detected about 80 percent of turbulence with at least a minute’s warning-enough time to seat passengers and flight attendants and clear aisles.