SLHIM (SLope Health Integrated Monitoring) is an intelligent, integrated system for monitoring slopes affected by landslides and assessing susceptibility to rainfall-induced landslides.
It is implemented integrating a hydrological-geotechnical (IHG) model, fed by sensor data, and a statistical analysis of rainfall thresholds.
IHG is a physically-based model that allows to assess the susceptibility of debris and earth slides of a few square kilometers (typically at 1:5,000 scale) thanks to a simplified soil-water balance and a stability analysis in effective stresses, based on the global Limit Equilibrium Method (LEM), possibly taking also into account uncertainties in the model parameters. IHG establishes a cause-effect relationship between rainfalls and site-specific groundwater level oscillation, and takes into account partially saturated conditions, in terms of additional resistive contribution, namely apparent cohesion, and in terms of estimated soil weight per unit volume.
The evaluation of the effectiveness of rainfall thresholds available for the Ligurian area is performed through the analysis of historical data of rainfall-induced landslides, representative of different rainfall regimes, with particular reference to extreme events (e.g. cloudbursts).The SLHIM system is based on the analysis of monitoring data (derived from on-site and/or remote sensor networks) over significant periods of time (i.e. months or years).
SLHIM, through the integration of software procedures, allows to process the data automatically or semi-automatically, identifying the state of activity of landslides and in-progress/potential slope instabilities thanks to a specific hydrological-geotechnical modelling.
The SLHIM system can be considered “intelligent” since it can assess the health condition of a slope. In fact, it is capable of checking the activity status of recognized landslides and assessing its attitude to reactivation or first trigger in case of rainfalls. Furthermore, sensor data and landslides susceptibility maps induced by rainfall, are shared on a geoportal and processed continuously through automated procedures and machine learning techniques. The previous large-scale landslide assessment enables the identification of one or more critical locations, which necessitate further attention to precisely quantify the risk of instability. This detailed analysis often requires the direct measurement of pore water capillary tensions inside partly saturated soils (negative pore water pressures), whose magnitude is directly related to the cohesive component of strength inside the ground. To this end, a sensor prototype is developed to continuously record the field variation of pore water capillary tensions caused by the interaction between the ground and the atmosphere. Various sensor designs are explored to maximize both recording range and stability of measurements over time. The recording range is assessed via evaporation tests while the stability of measurements is evaluated by monitoring the evolution of negative pore water pressure inside soil samples at the laboratory scale. In parallel, an automated data quality control algorithm is developed to assess the reliability of sensor measurements. The proposed algorithm compares, in real time, field data against a reference soil-water retention curve, which is calibrated for each potentially instable location, thus enabling a timely detection of possible sensor failures and/or malfunctioning.