Implementation of soil moisture data into landslide rainfall thresholds: two case studies in Italy and Norway

Prediction of rainfall-induced landslides is a complex task, due to the multitude of processes involved, heterogeneity of soil properties, spatial variability of rainfall and uncertainty in landslide inventories. Rainfall thresholds can provide a useful insight on the prediction of rainfall-induced landslides; however, they just describe a part of the problem, completely neglecting the hydrological conditions. Empirical thresholds, generally focus on the characteristics of precipitation, expressed in terms of intensity and duration (I-D threshold). Although an increasing number of studies is aiming at defining the link between precipitation characteristics and soil moisture data, few are describing the usefulness of soil moisture together with empirical thresholds for rainfall-induced landslide prediction. Soil moisture data are generally used in physically based models being a function of the characteristics of the soils therefore highly site-specific and obtainable with instrumental observations and/or in situ or laboratory analyzes.

In this study, a preliminary analysis on the use of soil moisture data for the definition of empirical rainfall thresholds is carried out. The newly released fifth-generation reanalysis product of the European Center for Medium Range Weather Forecasts (ECMWF), i.e., ERA5, provides soil moisture data even for those areas in which no measuring instruments are available. ERA5 data are available in the Climate Data Store on regular latitude-longitude grids at 0.1° x 0.1° resolution covering a period from 1950 to the present with hourly resolution. The goodness of the product has been verified comparing in situ available data with those obtained with ERA by statistical analysis including the Taylor diagram that links correlation coefficient, standard deviation and root mean squared difference between two analyzed series. Soil moisture data have been collected for several stations located in Norway and Italy.

Soil moisture data for Norway has been collected from stations in two different places near Oslo, while soil moisture data for Italy comes from the International Soil Moisture Network (ISMN), specifically, Calabria region stations have been used. Rainfall-soil moisture thresholds have been defined for two case studies and the performance of thresholds considering and neglecting the soil moisture has been evaluated.