Modelling possible structural instabilities of the Po River embankment, Italy, due to groundwater pumping in the Ferrara Province.

Land subsidence due to groundwater pumping is a major environmental problem in flat alluvial coastal plains. In the eastern Po River plain, Italy, CADF, a local Water Distribution Company serving a 1300 km2 territory, is producing water from a well-field established since the 1930s at the Ro Ferrarese village. The 107 m3/year current pumping rate is withdrawn from a 20-60 m deep confined aquifer. In order to cope with the increasing water demand, CADF has planned in 2006 the development a new field 5 km east of the old one. The new field is made of 9 wells located in the Po River overbank and produces from the same shallow aquifer. To provide the establishment authorization, the Regional Environmental Agency has required the evaluation of the environmental impact in terms of land subsidence and possible structural instability of the Po River embankment (that is more than 10 m above the surrounding farmland) due to differential displacements. In order to comply with the previous requirements, a modeling study supported by in situ investigations has been performed. Advanced three-dimensional finite element flow and poroelastic models have been implemented with a realistically detailed lithostratigraphic sequence obtained by integrating a number of borehole stratigraphies and 2D electrical resistivity tomographies. The models have been calibrated against the outcome of available pumping tests and records of land settlement taken at the old wellfield. A "best fit" scenario along with other three plausible scenarios have been simulated in order to account for the major data uncertainties of the hydrologic parameters, i.e. hydraulic conductivity of the pumped aquifer and sand and clay vertical compressibility. The models have been used to simulate the anthropogenic land subsidence from the establishment of the Ro well-field (1930) to the present time (2005), and to predict the expected occurrence up to 2035. The results show a cumulative maximum land settlement of 15 ± 4 cm at the Ro well-field over the simulated 105 years. A value of 6 ± 2 cm of land subsidence may be expected from 2005 to 2035 at the new well-field. The displacement gradient is of the order of 10-4 - 10-5. Hence, it may be concluded that the activation of the new well-field and the joint groundwater pumping from the old one should not generate any instability of the Po River embankment.