GNGTS 2016 - Atti del 35° Convegno Nazionale

250 GNGTS 2016 S essione 1.3 We show that, at the CDP, the abrupt and voluminous release of H 2 O-rich magmatic gases can heat hydrothermal fluids and rocks, triggering an accelerating deformation that can ultimately culminate in rock failure and eruption. We propose that magmas may be approaching the CDP at Campi Flegrei, a volcanic system in the metropolitan area of Naples, one of the most densely inhabited areas in the world. Large variations in the N 2 -He-CO 2 -Ar fumarole compositions, indicative of degassing in open-system conditions at increasingly lower pressures, are in fact accompanied by a 15 years long increase of the fumarolic carbon monoxide (CO) content, indicating a generalized heating up of the hydrothermal system. The N 2 /He ratio displayed, for example, a 20 years long decreasing trend that was interpreted as caused by decompression of a gas-magma separation process affecting a melt of primitive composition (i.e. trachybasalt, Caliro et al., 2014). Assuming typical redox conditions of hydrothermal systems, the measured increase of the CO/CO 2 ratio would correspond to a temperature increase from 210-220°C to 230-240°C (Chiodini et al. , 2015, 2011). Based on these observations Campi Flegrei magmatic system may be approaching the CDP, i.e. a depressurising magma batch is releasing fluids of progressively more H 2 O-rich (and CO 2 -poor) compositions, perturbing the thermal structure of the hydrothermal system. This hypothesis was tested and confirmed by comparing the observed geochemical changes with the results of TOUGH2 models of a hydrothermal system affected by the input of magmatic fluids, which increase their H 2 O/CO 2 ratio during time. The average temperatures of the system returned by the simulations, which were constrained by the CO 2 /H 2 O, CO 2 /CH 4 and N 2 /He fumarolic ratios, show in fact the same time evolution as that computed with the CO/CO 2 geothermometer. Furthermore the simulated temperature increase occur concurrently with the observed ground deformation and the increase in seismic activity. In the light of these results, we applied the Principal Component analysis (PCA) to the fumarole compositions to investigate the correlations among geochemical multivariate data with the geophysical signals acquired in the same period by the INGV monitoring systems. We started with the hypothesis that the geochemical dataset is correlated with few underlying processes, which characterize the state of the geothermal system, thus we are using the PCA to evaluate the number of processes and the possible correlation with the geophysical signals. Being geochemical data usually expressed in percentage we used the isometric log ratio to avoid the data collinearity. In the new space we evaluate the PCA using a robust algorithm which reject the outliers. ��� �������� ����������� ���� ���������� � ���������� ���� ����������� The unbiased statistical tool highlights a surprising high correlation among the independent data sets. Endogenous processes explain 90% of the variability of the gas composition and control ground deformation and earthquake occurrence. At Campi Flegrei an accelerating deformation is observed since 2005. The NeVoCGPS network (Neapolitan Volcanoes Continuous GPS) show a typical radial extensional pattern departing from the caldera centre (DeMartino et al ., 2014). Horizontal and vertical displacements follow in the time an exponential type curve with a maximum uplift of 40 cm reached in August 2016. The same temporal pattern of the ground displacements characterises the cumulative curve of the earthquakes when the proper statistic of the earthquake occurrence is considered. Campi Flegrei caldera is unlikely to be an isolated case. A similar, several-year-long period of accelerating inflation possibly driven by magma depressurization and heating processes was for example observed prior to eruptions at Rabaul, Papua New Guinea and Sierra Negra, Galapagos. Even if the magma underneath Campi Flegrei caldera is likely to be approaching the CDP, the possible future scenarios can be complicated by additional processes that have not been considered in our study. For example, increases in the melt liquidus due to H 2 O release and consequent magma crystallization could increase the melt viscosity, and therefore act against further magma migration. Additional careful scrutiny of monitoring data in the coming months and years is key to interpreting whether hydrothermal heating or magma quenching will prevail.