Development of innovative monitoring techniques and early-warning procedures for landslide and debris-flow hazard mitigation.

Debris flows are complex phenomena because they consist of a moving sediment–water mixture that evolves during runout through entrainment and deposition, causing rapid changes in concentration and density and producing unsteady, transient flow behaviour. Owing to this complexity, debris flows are difficult to reproduce in laboratory, and their initiation and propagation are difficult to capture by numerical models alone. Consequently, debris-flow research must rely on the documentation of real events through field monitoring in instrumented catchments. In this PhD project, multi-parameter monitoring data collected from several instrumented catchments were analysed to enhance understanding of debris-flow initiation, propagation, and internal dynamics. This thesis is co-financed by Hortus S.r.l. (Gallarate, Italy), and its scientific objectives were designed to ensure practical outcomes. The work is organised into three complementary studies, each addressing a distinct aspect of debris-flow processes. The first study focuses on debris-flow initiation and develops a method to define site-specific rainfall thresholds using short monitoring records, exploiting both triggering and non-triggering rainfall events. It relies on simple, low-cost instrumentation and can be transferred to catchments where reliable thresholds are needed within a few monitoring seasons. The second study investigates lahar dynamics using seismic recordings. The objective is to evaluate whether seismic signals can provide information beyond event detection by including frequency descriptors, and to assess how these metrics vary among flow types and within events. The third study uses multi-year monitoring data to examine debris-flow activity at the catchment scale, including event frequency and seasonality, transfer times from headwaters to the outlet, spatial variability of flow velocity along a channel reach, and the occurrence and characteristics of rapid secondary surges. Overall, this PhD project shows how integrated field monitoring and multi-sensor data analysis can strengthen process understanding and support the development of debris-flow warning strategies.