The influence of alluvial stratigraphic architecture on liquefaction phenomena: A case study from the Terre del Reno subsoil (southern Po plain, Italy)

Earthquake-induced liquefaction phenomena are commonly observed at shallow depths or at the ground surface above Holocene alluvial deposits. Nonetheless, a high-resolution examination of the relationship between surface manifestations and alluvial stratigraphy has been rarely attempted. In this study, we provide a detailed subsurface model from a selected sector of the southern Po Plain affected by widespread liquefaction processes triggered by the 2012 Emilia-Romagna seismic crisis. We report a comprehensive sedimentological and CPT-based characterization of the Terre del Reno subsoil and investigate how the stratigraphy controls earthquake-induced liquefaction phenomena. The main facies associations making up the alluvial subsoil constitute the basic elements of the liquefaction system, herein named host, source and sedimentary cap. Their subsurface distribution documents the presence of fluvial bodies prone to liquefaction beneath the earthen levees and across an area dominated by flat muddy floodplain deposits with no physical evidence of paleochannels at the surface. In addition, our observations suggest that vertical and lateral changes in stratigraphy control the mode and location of liquefaction evidence at the surface in both settings. In particular, the relative thickness of fluvial channel bodies vs. channel levee deposits (source-seal couplet) appears to control the severity of sand ejection at the surface. Our findings also suggest that vertical and lateral discontinuities within source layers prevent the occurrence of widespread liquefaction features. Conversely, lateral spreading effects and differential underground hydrostatic pressure at the interface between saturated and less permeable units are accountable for repeated sand ejection along surface fractures. This case study constitutes a unique natural laboratory thanks to the quality and density of available data. However, we emphasize that a similar approach can be useful to promote source-layer identification in areas characterized by poor or sparse data and to mitigate liquefaction hazards in urban areas built on alluvial plains.