The risk of a collision between aircraft is rising as the density of c
ommercial air traffic increases. This trend, together with the overwhe
lming need to upgrade the National Airspace System, has motivated the
Federal Aviation Administration to sponsor the development of metrics
to evaluate ''dynamic density'' - a proxy for the likelihood of collis
ion risk. Here we propose and evaluate a mathematical index of dynamic
density, D, that describes collision risk. Although our domain of inv
estigation is aviation, the logic of D is applicable whenever objects
move in limited spaces. A series of sensitivity analyses illustrate ho
w D responds to frequently encountered air traffic conflict situations
. We illustrate a use of D that characterizes pilot performance and ef
ficiency in experimental simulations of free flight and suggest other
human factors applications. This research could be applied immediately
by the traffic management units of en-route air traffic control cente
rs to reformulate the criterion for the critical capacity of sectors.