Modelling groundwater level fluctuations in rainfall-driven urbanized slow-moving landslides: first insights from case studies in the Liguria Region (NW Italy)

Slow-moving landslides are very slow or extremely slow landslides that often affect urbanized slopes, involving a wide range of soil and rock materials.  Often they can exhibit sudden changes in velocity related to local environmental changes, passing through slow (within 1 m/year) to rapid (more than 1 m/s) displacement. In built environments, the kinematic behaviour of these slope instabilities can lead to significant damage and even fatalities. Therefore, in active slow landslides, the prediction of movement acceleration is a crucial issue in the frame of landslide hazard and risk assessment for the design of warning systems and potential damage management. It is of great importance to investigate the factors that can drive velocity changes within unstable landslide bodies.

In this contribution, we focused on the role of hydrological preparatory and triggering factors (e.g., rainfall and groundwater level variations) on the unstable mass mobility. It is known that deep-seated slow-moving landslides are driven by pore-water pressure fluctuations that can result from infiltrating precipitation and/or snowmelt. However, the relationship between precipitation, hydrological responses and movement is not straightforward, primarily due to the complexity of the processes governing the recharge of groundwater in response to the rainfall regime, which can be influenced by many factors, both external (e.g., temperature, evapotranspiration, vegetation cover) and internal (e.g., layering, cracks, fissures). Therefore, including hydrological processes and their variability in landslide modelling is of paramount importance.

Here we present preliminary insights on the application of a simple physically-based model for quantifying groundwater fluctuations in response to discrete precipitation time-series in two reactivated slow-moving mass movements located in Liguria region (NW Italy): the Fontane landslide, in the Northern Apennines (eastern Liguria, Genoa Province) and the Mendatica landslide, in the Ligurian Alps (western Liguria, Imperia Province). Both landslides are rainfall-induced and affect small villages which have suffered damage in the past. The research activities are carried out in the framework of the RETURN (multi-Risk sciEnce for resilienT commUnities undeR a changiNg climate) project funded by the Italian MUR and the European Union Next-GenerationEU.

Long-term hydro-geotechnical monitoring data series (e.g., groundwater table levels) available for the two selected landslides and meteorological data (e.g., rainfall and temperature) from nearby measuring stations were collected and analyzed for two significant periods in order to grasp the seasonal fluctuations of the water table and the response to rainfall events. During the modelling, each period was split into two sub-periods: one, for the calibration phase, in which meteo-hydro-geotechnical data were used to estimate the parameters needed for the simulated water table to best approximate the measured one; the second, for the validation phase, in which the goodness of the model is verified. The outcomes of this study may represent an initial basis for gaining insights about the processes that influence groundwater table variations and defining models for the simulation/prediction of quantitative scenarios related to the hydrologic preparatory processes that influence the kinematic behaviour of the two selected slow-moving landslides. Indeed, the results of the groundwater model may be used as input data for predicting the landslide displacements.