LOCAL AND DISTRIBUTED HYDROLOGICAL AND SLOPE STABILITY MODELING FOR ASSESSING DEBRIS FLOW HAZARDS ALONG THE SARNO MOUNTAINS (SOUTHERN ITALY)

The carbonate mountain ranges bordering the Campanian plain (southern Italy) are mantled by ash-fall pyroclastic deposits, associated with eruptions of Mount Somma-Vesuvius and the Phlegraean Fields volcanic centers. The urbanized foot slopes along these mountains are among the most landslide-prone areas of Italy, and have been subjected to repeated occurrences of rainfall-induced debris flows. After the most recent catastrophic landsliding event occurred in the Sarno Mountains in May 1998, many studies focused on understanding the predisposing factors and triggering mechanisms, which are controlled by morphological and hydrological conditions, respectively. Debris flows in ash-fall pyroclastic soils are triggered during heavy rainfall events, especially if also preceded by prolonged rainy periods. However, the assumption that slope instabilities are triggered by a rainfall event whose magnitude equals or exceeds a threshold value does not provide reliable prediction of landslide occurrence. Rainfall patterns indirectly influence the stability of a slope through changes in pore pressures, while stability is also related to the spatial variability of geotechnical properties and of ash-fall pyroclastic cover thickness along slopes and road-cuts. The objective of this work is to advance distributed assessments of landslide hazards for the Sarno Mountains by accounting for known variations in the distribution of this cover thickness along slopes. Physically-based hydrologic modeling was used to derive deterministic rainfall intensity/duration thresholds for several specific landslide source areas. We present preliminary results from work to extend these locally derived thresholds into a distributed hazards assessment using the TRIGRS model. Different scenarios were simulated with TRIGRS to evaluate the stability of the ash-fall mantles and develop maps of landslide potential under different rainfall conditions. We demonstrate that this integration of numerical modeling, from local to regional scales, is a useful approach for mapping the spatial and temporal probability of landslide hazards that can account for the different hydrological antecedent conditions of the spatially variable ash-fall cover and temporally variable rainfall triggering.