Hydrogeochemical investigation of shallow aquifers before and after the 2012 Emilia seismic sequence (northern Italy)

Abstract

Soon after the MW = 6.0 main event of May 20, 2012 that struck the central part of the Emilia-Romagna region (Italy), several geochemical surveys were carried out on groundwater within the epicentral area. A total of 20 water samples were periodically collected, from May 2012 to July 2014, from shallow (up to 6 m depth) to deep (down to 175 m depth) wells within both unconfined and confined aquifers and analyzed for major ions, trace elements, dissolved gases and stable isotopes (δ2H–H2O, δ18O–H2O and δ13C-TDIC). Geochemical data were compared with previous data collected and analyzed in 2006 in a phase of absence of significant seismic activity. Monitored waters showed concentration variations in post-earthquake sampling on a large number of geochemical parameters. Many of these variations were recorded during the co-seismic phase and were transient as the geochemical parameters returned towards pre-earthquake values over time. The most significant transient variations involved trace elements, which generally show high sensitivity even to small variations in the surrounding environment due to their usually low concentrations in groundwater. Physical-chemical parameters (water temperature, pH, TDS) and major ions provided less unambiguous indications, whereas among dissolved gases CH4 and CO2 showed a general post-seismic increase within the unconfined aquifer. Increased contents of such gas species in both aquifers (this study) and soils (previous studies) suggest that the seismic-induced overpressure on the ground had enhancing effects on soil permeability and porosity and triggered their co-seismic migration upwards from deeper reservoirs. Water isotopes showed the systematic post-seismic change in δ2H toward heavier compositions with no significant changes in the δ18O, which was interpreted as due to isotopic exchange between water and a H2-bearing gas phase (e.g., H2S, CH4, H2) entering the system. Calculated δ13C data of CO2 suggested a relatively shallow production from both plant-root respiration and microbial-driven degradation of organic matter while the almost pure crustal origin of He (R/Ra values = 0.04–0.16 from the bubbling gas phase emitted by one of the monitored wells) reasonably excludes any evidence of both primary mantle 3He degassing and ascent of heavier CO2 from deep (mantle, decarbonation) inorganic sources. Monitored waters which showed the most significant transient variations are aligned in the same E-W direction along which the seismicity and soil gas anomalies were distributed, at about 5 km S from the epicenter of the May 20th seismic event and along the main direction of the May 29th (MW = 5.8) event. This confirms that the transient variations have been activated by the seismic sequence in a sector of the crust where the presence of a fault/fracture system favors the intensification of processes affecting sediments and groundwater (variations of porosity/permeability of soils, the groundwater level, redox state, etc.) and which are able to explain the observed geochemical variations. Only one sample monitored, the one closest to the epicenter of the May 20th event, showed clear geochemical evidence suggesting the hypothesis of mixing between superimposed water bodies.