[1] Carbon dioxide (CO 2 ) diffuse degassing structures (DDS) at Furnas volcano (São Miguel Island, Azores) are mostly associated with the main fumarolic fields, evidence that CO 2 soil degassing is the surface expression of rising steam from the hydrothermal system. Locations with anomalous CO 2 flux are mainly controlled by tectonic structures oriented WNW-ESE and NW-SE and by the geomorphology of the volcano, as evidenced by several DDS located in depressed areas associated with crater margins. Hydrothermal soil CO 2 emissions in Furnas volcano are estimated to be ∼968 t d . Discrimination between biogenic and hydrothermal CO 2 was determined using a statistical approach and the carbon isotope composition of the CO 2 efflux. Different sampling densities were used to evaluate uncertainty in the estimation of the total CO 2 flux and showed that a low density of points may not be adequate to quantify soil emanations from a relatively small DDS. Thermal energy release associated with diffuse degassing at Furnas caldera is about 118 MW (from an area of ∼4.8 km 2 ) based on the H 2 O/CO 2 ratio in fumarolic gas. The DDS also affect Furnas and Ribeira Quente villages, which are located inside the caldera and in the south flank of the volcano, respectively. At these sites, 58% and 98% of the houses are built over hydrothermal CO 2 emanations, and the populations are at risk due to potential high concentrations of CO 2 accumulating inside the dwellings.

No abstract

No abstract

Electrical conductivity tomography is a well-established galvanometric method for imaging the subsurface electrical conductivity distribution. We characterize the conductivity distribution of a set of volcanic structures that are different in terms of activity and morphology. For that purpose, we developed a large-scale inversion code named ECT-3D aimed at handling complex topographical effects like those encountered in volcanic areas. In addition, ECT-3D offers the possibility of using as input data the two components of the electrical field recorded at independent stations. Without prior information, a Gauss-Newton method with roughness constraints is used to solve the inverse problem. The roughening operator used to impose constraints is computed on unstructured tetrahedral elements to map complex geometries. We first benchmark ECT-3D on two synthetic tests. A first test using the topography of Mt. St Helens volcano (Washington, USA) demonstrates that we can successfully reconstruct the electrical conductivity field of an edifice marked by a strong topography and strong variations in the resistivity distribution. A second case study is used to demonstrate the versatility of the code in using the two components of the electrical field recorded on independent stations along the ground surface. Then, we apply our code to real data sets recorded at (i) a thermally active area of Yellowstone caldera (Wyoming, USA), (ii) a monogenetic dome on Furnas volcano (the Azores, Portugal), and (iii) the upper portion of the caldera of Kīlauea (Hawai'i, USA). The tomographies reveal some of the major structures of these volcanoes as well as identifying alteration associated with high surface conductivities. We also review the petrophysics underlying the interpretation of the electrical conductivity of fresh and altered volcanic rocks and molten rocks to show that electrical conductivity tomography cannot be used as a stand-alone technique due to the non-uniqueness in interpreting electrical conductivity tomograms. That said, new experimental data provide evidence regarding the strong role of alteration in the vicinity of preferential fluid flow paths including magmatic conduits and hydrothermal vents.

Time series of soil CO 2 efflux recorded in the Azores archipelago volcanic-hydrothermal areas feature daily and seasonal variations. The recorded CO 2 efflux values were lower during summer than in the winter season. The diurnal CO 2 efflux values were higher at dawn and lower in the early afternoon, contrary to that observed in biogenic environments. CO 2 efflux cycles correlated well with the environmental variables, such as air temperature, wind speed, and barometric pressure, which also showed low-and high-frequency periodicities. Several simulations were performed here using the Transport of Unsaturated Groundwater and Heat 2 (TOUGH2) geothermal simulator to complement the study of Rinaldi et al. (2012). The effects of the water table depth, air temperature perturbation amplitude, and soil thermal gradient contributed to an explanation of the contrasts observed in the diurnal (S 1 ) and semidiurnal (S 2 ) soil CO 2 efflux peaks for the different monitoring sites and seasons. Filtering techniques (multivariate regression analysis and fast Fourier transform filters) were also applied to the recorded time series to remove effects of external variables on the soil CO 2 efflux. The resulting time series (the residuals) correspond to the best approach to the deep-seated (volcanic/hydrothermal) CO 2 emissions and thus should be used in seismovolcanic monitoring programs. Even if no evident correlation can be established yet between the soil CO 2 residuals and seismicity over the monitored time, a seismic swarm that occurred around the end of 2008 might have triggered some deviations from the observed daily cycles.