Debris flow risk in the Dolomites

The Dolomites region (Eastern Alps of Italy) is a high-risk area of debris flow activity. Debris flows are triggered at the outlet of steep rocky, zero-order basins during intense convective storms; these watersheds quickly respond to rainfall and deliver a large amount of water at the head of steep channels filled by debris, where landslides are initiated by progressive channel bed erosion. Although this hazard is clearly recognized and the general distribution of potential active channels is well-known, the impact of these processes is becoming increasingly large. For instance, 14 major debris flows events occurred summer 2015 and three people were killed. Over the past few decades, in fact, the frequency of Locally Intense Convective Rainfall (LICR) has significantly increased because of the enhanced convection in the Mediterranean region. Consequently, both the frequency and magnitude of debris flows is getting higher. There is a need to better understand the hydro-meteorology forcing, and a need to develop better early-warning systems for the region. In this context, our research is focused on the early detection of critical rainfall (1) and to the evaluation of rainfall thresholds for debris flow occurrence (2). With regard to the first topic, one of the critical aspect of risk mitigation in the area is the warning time. At present, different techniques are used to raise an alarm (rainfall thresholds, meteorological radar, alarm systems) but the effectiveness of these methods is often vanished by the very short time available after the onset of the flow. We are trying to anticipate the alarm by measuring the time variation of atmospheric electric potential. Recent studies have shown that the arrival of a convective cell is preceded by a variation of the atmospheric electric field related to a particular lightning activity (ref). It seems possible to identify clear signs of an approaching thunderstorm already 1-3 hours before the start of precipitation, which is a valuable time for early warning purpose. As part of a previous research project (PRIN 2010), we installed an electric field sensor (Campbell CS110) in the source area of a debris flow in the Dolomites. The collected data look very promising in the perspective of anticipating the arrival of a critical rainfall. With regard to the second topic, the aim is to improve the empirical rainfall thresholds used to mitigate the debris flow hazard by combining field monitoring and numerical analysis. During previous research projects, we installed several monitoring systems to investigate the hydrologic response to rainfall in the initiation area of selected debris flow basins. We are now developing a simple numerical model for debris flow triggering to predict the critical runoff discharge in the initiation area, taking into account of the losses in the rocky watershed. This model will be deployed in the rainfall duration-intensity plane to obtain deterministic rainfall thresholds for debris flow triggering.