Spatial correlation analysis of near ship collision hotspots with local maritime traffic characteristics

A spatial correlation analysis of near collision clusters with local traffic characteristics is presented. The Moran's I and Getis-Ord Gi* spatial autocorrelation methods are used to determine whether near collisions show spatial clustering from global and local perspectives. The application of the developed approach to Automatic Identification System data of the maritime traffic off the coast of Portugal shows that there are several hotspots where the density of ship near collisions is relatively high. A co-occurrence analysis is then conducted to relate the near collision hotspots with selected local maritime traffic characteristics such as the average ship speed, speed dispersion, degree of speed acceleration, ship route overlaps and degree of angular deviation from ship route centreline. The identification of near collisions clusters and the assessment of the correlation of ship near collision hotspots with the maritime traffic features provide a means for improving maritime safety and reducing the occurrence of ship-ship collisions.