An event-based approach is developed and applied to 3 years of data of a four-link transhorizon path over the English Channel (La Manche). The signal enhancements (events) observed on the chart recordings during anomalous propagation are selected and, according to their dynamical signatures, categorized into three main patterns of ducting, layer reflection, and high k factor diffraction. The statistical analysis carried out on this event basis gives the correlation between the signal levels of different links and the cross correlation between signal levels and local N index derived from surface meteorological monitoring. The results reveal the fundamental properties of the anomalous propagation regimes. Using case studies of several interesting events, the meteorological processes and resultant impact on transhorizon transmission are also explained. This extensive work contributes to a better understanding of anomalous propagation mechanisms present in transhorizon links. ¿ American Geophysical Union 1995