What is it?

• A method to categorize a large set of trajectories into groups of similar trajectories

Goal is to:

• minimize differences among trajectories in a cluster
• maximize differences among the clusters

Results in:

• distinct clusters representing different synoptic regimes

Some applications:

• to estimate pollutant source regions
• to interpret forecast trajectory error
• to identify similar meteorological scenarios for case studies
• to compare AIRMoN precipitation data on a cluster-by-cluster basis