The continuous automatic monitoring in Tuscany (Italy) to identify possible hydrogeochemical precursors of earthquakes

Since late 2002, in the framework of a project funded by Seismic Prevention Sector of Tuscany Region, the IGG-CNR Pisa has setup the "Tuscany Geochemistry Network" to identify possible hydrogeochemical precursors of earthquakes. For this purpose six continuous automatic monitoring stations were designed, built and installed in some areas of highest seismic risk in Tuscany (Garfagnana, Lunigiana, Mugello, Tiber Valley and Amiata Mount) and a systematic study of the hydrogeochemical characteristics, both chemical (Na, K, Ca, Mg, Cl, SO4, SiO2, F, B), isotopical (d18O, dD, Trizio) and chemico-physical (T, pH, CE) has been carried out. The continuous automatic monitoring stations operate with flowing water (about 5 litres per minute) and record the following parameters: temperature (T), pH, electrical conductivity (EC), redox potential (ORP) and the content of CO2 and CH4 dissolved in water. For CO2 and CH4 determination a cell extraction of the gases dissolved in water has been specifically designed. Frequency of data acquisition is 1 second for all parameters; the average, median and variance values are locally calculated over a period of 5 min and recorded with a local data logger. Data are finally transmitted once a day to CNR-IGG by router GPRS and available for processing and interpretation. A geological and hydro-structural characterization of the Garfagnana and Lunigiana areas has also been integrated, to define a bidimensional model of the circulation, to reconstruct the geometry of the aquifers, the geometry of the fault systems and the emergency typology of the deep circulation. The results of almost ten years of continuous monitoring in Tuscany can be summarised as follows: i) the monitoring stations made possible to detect even small variations of important parameters possibly related to seismic activity. ii) acquired data allowed to characterize the groundwater circulation patterns; iii) in all locations, the observed trend of the automatically acquired parameters is consistent with the calibrating periodic manual sampling results. iiii) When a hydro-geochemical study is integrate with a hydro-structural model (e.g. Lunigiana and Garfagnana), the overall study can better characterize the significance of detected anomalies, disentangling variations in parameters due to natural phenomena (recharge from specific areas, mixing processes...) from variations possibly due to change in permeability of the system and therefore related in some way to seismic activity. In these ten years of monitoring several significative variations have been recorded before earthquakes.