Historical floods and a recent extreme rainfall event in the Murgia karstic environment (Southern Italy)

In karstic areas, the mean annual stream flow is low in terms of mean annual net rainfall percentage, and, in many cases, stream flow is so rare to be considered an exceptional independent of the peak flow value. Nevertheless, many types of flooding processes can be identified in karsts. We consider the recharge-related flooding of a bare karst, with a case study of a Mediterranean area dominated by semi-arid climate. In these cases floods are rare, short, and sudden but the effects are often very dramatic. Three types of analyses were used. The historical analysis focused on the series of past or historical flood events, which provided a means to define the flood recurrence period and damage levels, and the main steps of anthropogenic development expressed by the population number and needs. The hydrological analysis considered long time series of annual maximum short rainfalls, in order to assess the return period of high intensity rainfall that caused the studied flood event. The geomorphological analysis considered rainfall, the hydrogeological characteristics of outcropping rocks and soils and the geomorphological data of the drainage basin to calculate net rainfall and peak flow characteristics. We explored historical flood events for the study area, characterising the role of human effects on damages, and used a recent flood event to test our capability to estimate ephemeral river flow on the basis of rainfall measurements. The historical analysis of floods and damages highlighted four flood periods, distinguished on the basis of the recurrence period and the level of damages. It is clear that the lessons of past flood events are quickly forgotten. The hydro-geomorphological analyses showed the peculiarities of the 2005 flood event. The maximum rainfall return period for the 2005 event was due to exceptional three-hour cumulative rainfall. This study shows that we can estimate peak flow characteristics even when time series of river flows are not available, thereby allowing us to improve our capability to forecast the flooding effects of severe storms in karstic areas.