A mathematical morphology approach for a qualitative exploration of drought events in space and time

Drought events occur worldwide and possibly incur severe consequences. Trying to understand and characterize drought events is of considerable importance in order to improve the preparedness for coping with future events. In this paper, we present a methodology that allows for the delineation of drought events by exploiting their spatiotemporal nature. To that end, we apply operators borrowed from mathematical morphology to represent drought events as connected components in space and time. As an illustration, we identify drought events on the basis of a 35-year data set of daily soil moisture values covering mainland Australia. We then extract characteristics reflecting the affected area, duration and intensity from the proposed representation of a drought event in order to illustrate the impact of tuning parameters in the methodology presented. Yet, this paper we refrain from comparing with other drought delineation methods.