On the use of inverse analysis for the estimation of soil hydraulic and retention parameters from monitoring data of a river embankment

Seepage in the vadose zone greatly depends on the hydraulic properties of the soil. Their experimental determination often requires specific equipment and time-consuming procedures; furthermore, in many applications, such as river embankments, the spatial variability of these properties turns out to be particularly significant. Although the use of a monitoring system is often an essential aspect of engineering projects, quantitative tools for the calibration of predictive models based on site measurements are not fully explored Recently, indirect estimate of parameters through inverse analysis has proven to be a valid alternative to direct methods (i.e. site and laboratory investigations), especially in large-scale studies, in which an extensive set of input data needs to be defined. The present study proposes an in-depth investigation of limitations and potentials of inverse modelling procedures for the estimation of soil hydraulic parameters. The case study is a multi-layered river embankment section along the river Secchia, a right-hand tributary of the river Po (Northern Italy), subjected to transient seepage phenomena in saturated and unsaturated conditions due to river level fluctuations and soil-atmosphere interaction. Site monitoring of water content and pore water pressure within the riverbank body ensures a wide set of observation data to be used for the calibration of a FE flow model. Numerical simulations using different retention models combined with the local optimization method based on the Levenberg Marquardt algorithm have been considered to perform the inverse analysis. Finally, a comparison among indirect simulation performances is presented through qualitative and quantitative methodologies.