I am measuring seasonal peaks in the occurrence of a natural phenomenon. Some areas show a single peak in activity each year. Others have two peaks separated by a trough. I am looking for a way to quantify the degree to which a distribution is bi-modal.