(Campi Flegrei Caldera, southern Italy) from active

Journal of Volcanology and Geothermal Research 357 (2018) 177–185

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research journal homepage: www.elsevier.com/locate/jvolgeores

High resolution, multi-2D seismic imaging of Solfatara crater (Campi Flegrei Caldera, southern Italy) from active seismic data S. Gammaldi a,⁎, O. Amoroso a,1, L. D'Auria b, A. Zollo a a b

Department of Physics “E. Pancini”, Università Degli Studi di Napoli “Federico II”, Napoli, Italy Instituto Volcanológico de Canarias, Tenerife, Spain

a r t i c l e

i n f o

Article history: Received 19 July 2017 Received in revised form 20 March 2018 Accepted 28 March 2018 Available online 27 April 2018

a b s t r a c t A multi-2D imaging of the Solfatara Crater inside the Campi Flegrei Caldera, was obtained by the joint interpretation of geophysical evidences and the new active seismic dataset acquired during the RICEN experiment (EU project MEDSUV) in 2014. We used a total of 17,894 first P-wave arrival times manually picked on pre-processed waveforms, recorded along two 1D profiles criss-crossing the inner Solfatara crater, and performed a tomographic inversion based on a multi-scale strategy and a Bayesian estimation of velocity parameters. The resulting tomographic images provide evidence for a low velocity (500–1500 m/s) water saturated deeper layer at West near the outcropping evidence of the Fangaia, contrasted by a high velocity (2000–3200 m/s) layer correlated with a consolidated tephra deposit. The transition velocity range (1500–2000 m/s) layer suggests a possible presence of a gas-rich, accumulation volume. Thanks to the mutual P-wave velocity model, we infer a detailed image for the gas migration path to the Earth surface. The gasses coming from the deep hydrothermal plume accumulate in the central and most depressed area of the Solfatara being trapped by the meteoric water saturated layer. Therefore, the gasses are transmitted through the buried fault toward the east part of the crater, where the ring faults facilitate the release as confirmed by the fumaroles. Starting from the eastern surface evidence of the gas releasing in the Bocca Grande and Bocca Nuova fumaroles, and the presence of the central deeper plume we suggest a fault situated in the central part of the crater which seems to represent the main buried conduit among them plays a key role. © 2018 Elsevier B.V. All rights reserved.

1. Introduction The Campi Flegrei (hereinafter Cfc) caldera is a volcanic complex located west of Naples, the most populated city in southern Italy. It extends over an area of about 400 km2 and originates from a nested resurgent caldera resulting from two large collapses associated to the Campanian Ignimbrite (37 ky) and the Neapolitan Yellow Tuff (19 ky) eruptions (Orsi et al., 1996). The last eruption in the caldera occurred in 1538 (Mt. Nuovo Eruption) while its more recent activity (1972 to present time) is characterized by large ground deformation, seismicity associated to uplift episodes (bradyseism), intense and diffuse degassing and fumarolic emissions. The caldera and its densely populated surroundings constitute one of the areas with the highest volcanic risk of the world. Kilburn et al. (2017) have just discovered that, the combined corrected uplift at Cfc (with intervals of stasis removed), also follows the classic elastic-brittle sequence for deformation in extension. The crust, they say, within uplift of about 4.2 m suggests that it is ⁎ Corresponding author. E-mail address: [email protected] (S. Gammaldi). 1 Now at Department of Physics “E. R. Caianiello”, Università degli Studi di Salerno, Fisciano, SA, Italy.

https://doi.org/10.1016/j.jvolgeores.2018.03.025 0377-0273/© 2018 Elsevier B.V. All rights reserved.

now approaching the transition from quasi-elastic to inelastic deformation. The Solfatara volcano, one the main monogenic craters of the Cfc, is characterized by an intense shallow hydrothermal activity, due to the interaction between a convective plume in the hydrothermal system of Cfc and meteoric inputs which produce seasonal changes in the physical properties of the area (Caliro et al., 2007). A more accurate knowledge about the hydrothermal activity allows understanding of dynamics of the hydrothermal system of the caldera and its relationship with possible pre-eruptive processes. The hydrothermal activity of Cfc is also a major factor in driving the seismicity of Campi Flegrei (D'Auria et al., 2011). Moreover, Chiodini et al. (2017) studied the 2000–2016 CFc background seismicity, mainly characterized by swarm-type occurrence of low-magnitude volcanic quakes, correlated with the ground deformation since the 2008, are basically identical. The cumulative distribution, displays a similar positive correlation with the fumarolic CO/CO2 ratio (i.e. Fig. 4b, d in Chiodini et al. (2017)) that is the most sensitive gasgeothermometer for hydrothermal systems. The first evidence in the last 30 years, of a renewed magmatic activity at Campi Flegrei after the unrest phase began in 2005, is given by D'Auria et al. (2015). They correlate the simultaneous occurrence of

178

S. Gammaldi et al. / Journal of Volcanology and Geothermal Research 357 (2018) 177–185

the swarm which occurred on September 7th, 2012 beneath the town of Pozzuoli, and the first magma injection rate peak, which suggests a strong relationship between the two phenomena. In particular, in this work, we want to highlight the structure and dynamics of Solfatara crater which is within the area having the highest probability of vent opening in case of renewal of volcanism in shortmidterm in the Campi Flegrei volcanic and seismic hazard assessment (Orsi et al., 2004; Convertito and Zollo, 2011). High resolution imaging, through the joint interpretation of seismic velocity and attenuation tomographic images, together with geological and other geophysical evidences has shown to be a useful tool to study areas characterized by a complex tectonic setting (Di Stefano et al., 2011, Amoroso et al., 2014, 2017a), geothermal (Gunasekera et al., 2003; Amoroso et al., 2017b) and volcanic areas (Pasquet et al., 2016; De Landro et al., 2017; Serlenga et al., 2016) allowing mapping the presence of fluids. Repeated seismic tomography (i.e. 4D tomography) is used to track the fluid movements involved in oil exploration and reservoir and volcano monitoring (Lumley, 2001; Patanè et al., 2006) as well as for the comprehension of the role of the fluids on faulting mechanisms in the tectonic environment (Chiarabba et al., 2009). To assess the large variety of non-linear thermo-mechanical and chemical processes associated with hydrothermal systems, which precede and accompany unrest episodes, the Solfatara crater was chosen as one of the case studies of the Mediterranean Supersite Volcanoes (MED-SUV) project. During this project field activities, lab-scale experiments, processing and modelling issues were performed in order to determine the internal structure of Solfatara. In particular to provide highresolution images of the structure beneath the Solfatara, and to detect and track changes in the medium, the “Repeated Induced Earthquake and Noise” (RICEN) seismic experiment was designed and carried out. During this experiment seismic data coming from passive and active sources with different array geometries were acquired. They were later combined with resistivity, seismo-electric and magnetotelluric measurements. In particular, using the 3D geometries acquired during the campaign, De Landro et al. (2017) performing the ultra-high-resolution P-wave imaging identified the gas rich area using a linear inversion method up to 35 m in depth. Moreover, the shallowest area of the crater was imaged by Serra et al. (2016) through the surface waves. The authors evidenced a strong heterogeneous hydrothermal area due to the presence of a water layer close to the Fangaia area and an abrupt variation in the NE direction. More recently Bruno et al. (2017) obtained a seismic image of the deeper reflectors of the Solfatara Crater using the same dataset of this work. In addition to these previous studies, in this work we realized two tomographic images of P-wave velocity, relative to the 2D seismic profiles realized during two different phases of the RICEN experiment. We used a non-linear travel-time tomographic method based on the evaluation of the “a posteriori” probability density through the Bayesian approach of Zollo et al. (2002). Adopting a multiscale technique, we obtained higher resolution and major penetration depth images of the shallow structure inside the volcano Solfatara compared to the aforementioned works. The two 2D tomographic images, were jointly interpreted with geological and structural information, providing new insights into shallow fluid circulation system. 2. Solfatara geological settings The Solfatara volcano (hereinafter Sv) (Di Vito et al., 1999; Isaia et al., 2009), located in the North-East of the town of Pozzuoli, has a diameter of 0.6 km and an average elevation of 180 m above sea level. The tuff cone was formed 3815 ± 55 y b.p., during the third epoch of volcanic activity of the CFc, as evidenced by stratigraphic and geochronological data (Di Vito et al., 1999). Daily the volcano releases large amounts of heat through the fumaroles inside the caldera (Bocca Grande and Bocca Piccola), with an estimated CO2 flux of about 460 tons/day (Aiuppa et al., 2013) and through diffuse degassing with an estimated

rate of 1500 tons/day CO2 (Chiodini et al., 2001). Several authors have proposed volcanic models to link the fumarolic activities with the buried geological structures. Petrosino et al. (2012), using geophysical and geochemical data, proposed an evolutionary model of the Sv. In their work, the volcano has been characterized by a long-lasting multievent activity, due to vent migration along structurally controlled fissures. The identified the geometry of two tuff-cone craters: a former crater and a younger nested crater. The east part of the volcano presents the tephra associated with the younger eruption (consolidated tephra) overlaid with a thinner reworked deposit. Westward, the tephra of the younger crater disappears, and is replaced by the reworked deposit covering a lava dome. In interpreting the electrical resistivity tomography carried out within the Sv, Byrdina et al. (2014) reconstructed the vadose zone, the aquifers, the gas-rich bodies and the mixture areas. Especially in the E-W profiles, they identified a low resistivity area below the Fangaia associated with the shallower aquifer and the underlying plume, while the high resistivity area in the eastern part was interpreted as gas filled bodies. Furthermore, the presence of these bodies is confirmed by a very recent 3D ultra-high-resolution P-wave imaging obtained by De Landro et al. (2017). They highlighted the presence of a high P-wave velocity anomaly (N1500 m/s) eastward interpreted as a volume of gas accumulation, counterposed to a low velocity anomaly associated with the aquifer. The volcano structure has been also identified as a maar diatreme structure, a system in which there are several faults and fissures (Isaia et al., 2015). We reported in Fig. 1 the WNW-ESE buried fault trace as previously mapped by Vitale and Isaia (2014). We noted that in Isaia et al. (2015) the buried fault splits the areas into two: one with low and another with high resistivity value. They were interpreted postulating different fluid contents: the low resistivity area is associated with the shallower meteoric water (Fangaia) overlapped to the deeper hydrothermal plume, contrasted by the high resistivity anomaly correlated with gas in the eastern area. 3. Data, method and inversion strategy 3.1. The RICEN experiment In the framework of the seismic survey Repeated InduCed Earthquakes and Noise (RICEN), new high-quality data has been made available for the study and the understanding of this complex volcanic system (MED-SUV project, http://med-suv.eu). The RICEN experiment target mainly consisted in studying the variations of a volcanic structure, through the detection of space-time changes in seismic records acquired at many sensors (Festa et al., 2015; Gammaldi, Ph.D. thesis, 2018). With this aim, during the experiment, the seismic waves (both in passive and active acquisition modes) were used as a diagnostic tool for 3D and 4D seismic imaging. Moreover, several resistivity, seismo-electric and magneto-telluric soundings were carried out (Gresse et al., 2017; Amoroso et al., 2017b). The RICEN experiment was organized into three different phases. In the first phase (RICEN1), the feasibility of a repeated experiment and the resolution of the resulting velocity model were studied to observe the size and the relevance of the anomalies to be recognized at the Solfatara. The pilot phase consisted of the joint acquisition, from both sparsely distributed three components geophones inside the crater, and seismic stations placed on a regular grid of 115 × 90 m2 area (named 3D experiment) in front of the Fangaia (Fig. 1). In the second phase (RICEN2) besides the two joint acquisitions mentioned previously, hosted a 2D profile, with NNE–SSW direction, which was performed on May 21, 2014. The same configuration was used for the last phase (RICEN3) with the difference that the 2D profile, performed on November 11, was oriented orthogonally to the one performed during the RICEN first act. The seismic signals have been produced by a Vibroseis truck. The truck is a 6400 kg IVI-MINIVIB ® which a theoretical pick force of ~27 kN at each sweep with a frequency range of 5–150 Hz (Bruno et al., 2017).


179

Fig. 1. Solfatara crater location and RICEN experiment acquisition layout. In the embedded upper box, we represent the position of the Campi Flegrei caldera respect to Italy on the left side; the shaded relief of the two calderas and crater rims of CFc are on the right side. In the main box the Solfatara crater is shown with the arrays positions, the Fangaia area and the main fumaroles position. The symbol legend is in the lower box. The map in this figure has been obtained with Google Earth 7.1.8.3036 2017: Solfatara, Pozzuoli, Metropolitan City of Naples, Italy retrieved from https://www.google.com/maps/@40. 82802,14.13932,17z/data=!3m1!1e3.

3.2. Data The signals have been sampled with a frequency rate of 1000 Hz. The first array was about 430 m long and oriented NNW-SSE while the second was about 480 m long and oriented along a WNW-ESE direction (Fig. 1). The interval between receiver positions was 2 m for a total of 240 receiver points along the profile A and 216 along the profile B. The sources were split-spread deployed with a source-interval average distance of 4 m. In total, 116 and 75 sources acquired data in May and in November respectively. We collected the records in Common Shot Gather (CSG) and displayed the signal relative to a shot acquired by the receivers. To highlight the first arrival times in the recorded signal, we applied a specific strategy of data processing. The raw dataset was given by convolution of the instrumental response (Serra et al., 2016) by the source time function and the Green's function of the medium. The source time function is the sweep whose frequency increases linearly with the time. To obtain the source-corrected record, we needed to deconvolve the signal to remove the source time function. Usually, this is performed in frequency domain but according to Serra et al. (2016) we operated in time domain by the cross-correlation of the sweep with the observed signals. Since, for a specific source time function the associated autocorrelation is a well-known wavelet, for a linear sweep the autocorrelation of the source is the Klauder wavelet

(Robinson and Saggaf, 2001). Hence, the original signal was cross-correlated with the sweep generated retrieving the source-corrected signal filtered by the Klauder wavelet. Moreover, to preserve causality in the data, an additional minimum phase filter was applied after the crosscorrelation (Gibson and Larner, 1984). The deconvolved sourcecorrected record did not allow us to detect the first arrival time yet, because of the amplitude saturation effect due to late, large amplitude ground roll (Serra et al., 2016). To enhance the amplitude of the direct and refracted waves, we applied an automatic gain control (AGC) with a mobile window of 200 ms with a scale value amplitude equal to 1. Finally, we manually picked the first arrival time by visual inspection of each section obtaining a total amount of 9053 and 8841 picks for each array (Fig. 2). Furthermore, we performed a preliminary 1D velocity analysis of Vp for each CSG. This analysis consisted in generating a starting model from the intercept-time of the refraction onsets (Fig. 2). In the Supplementary material, we show an example of 1D velocity profile obtained for the shot 35 (Fig. 1). 3.3. Method and inversion strategy The inference techniques are more difficult in the presence of strong lateral heterogeneity of the velocity field due to complexity of volcanic structures which increases the difficulties in the modelling. High

180


Fig. 2. Processing applied on the CSG associated the shot 35 (see position in Fig. 1) and travel-time plot of the whole dataset. a) Example of CSG section relative to shot 35 b) the same CSG after AGC processing (c) travel-time relative to the Model A d) travel-time relative to the Model B.

velocity anomalies have been observed in shallow crustal regions below numerous active volcanoes (Lees, 1992; Benz et al., 1996; Okubo et al., 1997; Villasenor et al., 1998; Laigle and Hirn, 1999; Chiarabba et al., 2000). To reduce the dependence of the final tomographic solution on the initial model, we used a non-linear travel time tomography approach (Zollo et al., 2002). The method consists in retrieving the velocity model through the evaluation of the maximum likelihood model by maximizing an “a posteriori” probability density function. The optimization can be performed by the sequential use of the Genetic Algorithm and the Simplex (Press et al., 1986; Goldberg, 1989; Whitley, 1994). The travel times used in the forward modelling are computed using a classical two-point raytracing approach based on a shooting technique (Rawlinson et al., 2007). The 2D velocity field is parametrized through a regular grid of nodes. The velocity at any point of a continuous medium is computed by 2D cubic spline interpolation function (Virieux, 1991). The inversion strategy was based on a sequential multiscale approach (Chiao and Kuo, 2001). This approach consisted in a series of inversions, performed at progressively lower grid spacing. In our case the best parametrization model corresponds to the 9 × 5 grid nodes according to Akaike (1974) shown in Table 1. The actual size of the model was chosen accordingly with the length of the array (Fig. 1) and with the maximum depth reached by the rays. The parameterizations involved in the multiscale were the same for the two analysed profiles and all the details about the number of nodes and grid spacing are resumed in Table 1. We assessed the model quality solution and the associated uncertainties based on the evaluation of the conditional probability density function (PDF) at each node of the model and on ray coverage (Zollo et al., 2002). Furthermore, to assess the geometry retrieved in the model we perform several “fixed geometry tests” (see Supplementary material), while to check the travel time used in the inversion a data

quality test was performed. More details on the tests can be found in the supporting information (see Supplementary material). 4. Results In Fig. 3a and c we show the travel-time residuals, calculated for the model with the best fit parameterization, as a function of the source-receiver distance and their occurrence (Fig. 3b and d) for A and B profiles. The average and the standard deviation are −0.2 ms and 8.2 ms respectively for the model A, and 0.2 ms and 4.6 ms for the model B with a final misfit reduction equal to 52% and 58%, respectively. According with the PDF calculated at each node of the grid and to the ray coverage (Supplementary material), the well resolved areas of the models have a maximum depth of 60 m, a nodes resolution spacing of 56.25 and 53.75 m in horizontal respectively for the Model A and B and 22,5 m in vertical for both the models. Furthermore, in Fig. 4, we show a matching of the 2D velocity models obtained independently, plotting a 1D profile at the intersection Table 1 Residual distribution in terms of media and standard deviation, grid spacing in meters, AIC value, for all the model parameterization for both the models A and B. Nodes (nr)

μ (ms)

σ (ms)

Grid spacing (m)

AIC

Model A I II III IV

5×2 5×3 9×5 17 × 9

1.7 1 −0.2 −0.3

12.3 11.4 8.2 7.9

112,5 × 90 112,5 × 45 56,25 × 22,5 X

96 93 90 111

Model B I II III IV

5×2 5×3 9×5 17 × 9

3.1 3.1 −0.2 −0.2

11 9.2 4.6 4.3

107,5 × 90 107,5 × 45 53,75 × 22,5 X

80 66 63 85


181

Fig. 3. Final residuals plot for model A (top) and model B (bottom). a) The residuals vs source-receiver distances for model A; b) the residuals occurrence distribution histogram for the model A; c) residuals vs source-receiver distances for model B; d) the residuals occurrence distribution histogram for the model B.

point. It is possible to visualize the correspondence of the velocity values at the crossing points of the two profiles, with a maximum velocity difference of 500 m/s at the maximum resolved depth. From the two tomographic images reported in Fig. 4 we can identify: a P-wave velocity increase with the depth, a very low velocity layer down to 500 m/s in the first 15 m; a low velocity layer with values ranging between 500 and 1500 m/s at 20–40 m depth; a medium velocity zone with values of 1500–2000 m/s at 30–60 m depth; a high velocity layer with values ranging from 2000 to 3200 m/s at 40–60 m depth. The main anomalies and velocity range were tested by the use of the fixed geometry test. In the latter we perform the method with the real acquisition geometries. Concerning the lateral variations, the two models exhibit different characteristics. The profile A presents a progressive eastward velocity transition, at offset 220–280 m, which divides the model into two regions: at East a high velocity region, with corresponding Vp range of 2000–2800 m/s, and at West a low velocity region (Vp ranging in

1000–2000 m/s). The profile B is characterized by a low Vp anomaly, with values ranging between 1500 and 2000 m/s, located at 200–250 m offset buried in a layered pattern.

5. Discussions The sharp velocity transition at offset 220–280 m, characterizing the model A, is correlated with the decrease of the thickness of the unconsolidated tephra replaced by the increase of the velocity gradient moving eastward with Vp values ranging of 2000–2800 m/s (Fig. 5). These high velocity values indicate the presence of consolidated tuff formation at east associated to the younger nested crater (Petrosino et al., 2012). The two formations, characterized by variable consolidation, are associated with different fluids percolation: the low velocity range at west is mostly interpreted with the hydrothermal plume of Fangaia and, the eastern high velocity anomaly with the presence of a gas plume.

182


Fig. 4. Vp tomographic results. a) Model A relative to the best parameterization grid acquired during WNW-ESE oriented profile, while in b) model B relative to NNW-SSE oriented profile. c) The 1D Vp profiles extracted from the two tomographic models close to the intersection point the circles are the profiles of model A and the triangles are profiles of model B.

Moreover, the fluids contrast is represented by low resistivity values at west and high values at east provided by Byrdina et al. (2014). The detailed volcanic formation and the fluids content, led us to provide a new insight about the connection between the shallow hydrothermal activity and its surface evidence in the BG and BN fumaroles of the Solfatara crater. We distinguish the low velocity zone, as saturated with meteoric water located in the most depressed area of the crater (Fangaia), overlaid by the higher velocity zone correlated with the lava dome (Isaia et al., 2015) which represents a cap rock for the underlying deeper plume. Nevertheless, there are the presence of the cap rocks at west, represented by the lava dome, and at east, represented by the unconsolidated tephra, the gasses of this plume are released in the eastern part of Sv, as testified by the presence of the vigorous fumaroles BG and BN. To better understand this system, we must discuss the high velocity anomaly in model B. The latter is associated with the consolidated tephra (Fig. 5), also visible in model A, which is, in the model B, separated by a low Vp anomaly, within a range of 1500–2000 m/s. This

anomaly perfectly matches the position of the fault mapped in Fig. 1 (Vitale and Isaia, 2014). This fault plays a key role in the local system of the fluids circulation. In fact, the deep hydrothermal plume (Caliro et al., 2007; Zollo et al., 2008; De Siena et al., 2010) is a mix of water and volatile products, whose gas migrates through a “buried conduit” and being finally released through BG and BN. Consequently, the fault mapped on our seismic tomography analysis seems to be a preferential gas migration path within these mostly impermeable rocks characterizing the crater. Moreover, the fault plane has a depth extension of at least 90 m starting from the deeper part of the crater (source of the hydrothermal plume), corresponding to sea level (Isaia et al., 2015), up to the more recent deposits. Below this recent deposit, an independent tomographic study of De Landro et al. (2017) provided the evidence for a gas accumulating volume, whose velocity and geometry well matches with our model. Moreover, our results show that this volume dips toward the Fangaia and becomes thinner in the eastern shallower part according to the presence of the high velocity zone, i.e. a consolidated tephra of the younger crater.

Fig. 5. Interpretation of the two Vp velocity model: at the bottom WNW-ESE the sketch relative to the model A, to the bottom the NNE-SSW sketch relative to the model B. In the right side the table of the geological interpretation of the velocity range.


The high-resolution tomographic images led us to map the shallower gas migration path through the joint interpretation of 2D models of A and B profiles. The multi-2D imaging in Fig. 6 shows a new constraint that we infer for the Solfatara system crater. The fault mapped in Fig. 1 represents the main conduit between the deeper hydrothermal source (indicated as plume) and the main fumaroles of BG and BN. The reason for the bordering degassing system respect to the central gas accumulation, is testified by third cap rock role represented by western unconsolidated water saturated tephra, as indicated in Figs. 5 and 6 with Vp velocity values up to 1500 m/s. To sum up we propose a new constraint for the shallower fluid circulation system through a driving buried fault, trapped by a cap rock. The gas raises the shallower central area of the crater through the fault up to 40 m in depth. Here the ascent of the gas is blocked by the shallower cap rock represented by the unconsolidated tephra which is saturated because it is in the most depressed area of the crater surface where the water accumulates. In the eastern part of the crater, indeed, the saturated tephra is thinner because leaned forward to the depression. Therefore, the water cannot stagnate, allowing the underlying gas plume to escape. In fact in this area, between 200 and 300 m offset in the model A, we have clear evidence of gas accumulation, with more accurate details, and constrained by our tomographic analysis and criss-crossing interpretation. In model B, indeed, this area corresponds to the low velocity anomaly crossed by the fault. Observing the velocity changes going from western part of the A model to the eastern area, the gas accumulation volume seems to disappear. Finally, in this part there are no more conditions for the accumulation of gas, because of the lithological formations whose layering, leaning and composition contrasts drive the fluids toward the east part of the crater, where the ring faults facilitate the releasing demonstrated by the fumaroles of BG and BN.

183

6. Conclusions In this work, we obtained 2D high-resolution images of P-wave velocity beneath the Sv by applying a non-linear inversion strategy based on a multi-scale progressive approach. The images cross the volcanic crater along two profiles that have been interpreted jointly with geophysical evidence coming from this work and from previous studies. By the joint interpretation of geophysical evidence and the new two tomographic images, we provided the shallower stratigraphy of the Sv in terms of the P-wave velocity range. Proceeding from the top to the bottom of the crater we recognize the following stratigraphic sequence: the recent deposit with a Vp up to 500 m/s, the unconsolidated tephra with a Vp between 500 and 1500 m/s which is saturated in water, the gas migration path with a Vp between 1500 and 2000 m/s and finally, below, the consolidated tephra and the lava dome with Vp velocity range from 2000 m/s up to a maximum of velocity retrieved of 3262 m/s. By multi-2D seismic imaging we were able to determine a new shallower fluid circulation system. The origin of the gasses released in BG and BN fumaroles, located at the border of the crater, is associated with a deeper plume located at the centre of the crater. Through the high-resolution multi-2D tomographic images retrieved in this work, we provide a more detailed picture of the gas migration path through two main structures. The central fault represents the highpermeability pathway for hydrothermal fluids. The gasses at depths of 40–60 m are blocked by the tephra saturated in meteoric water, which is the main shallower cap rock. Starting from the middle of the crater going westward, the saturated tephra became thinner being replaced by the gas accumulation zone. The gas is channelled between the consolidated and unconsolidated tephra, and finally released by the ring faults bordering the Solfatara. The latter

Fig. 6. Multi 2D sketch: Join interpretation of the two models with the fault mapped in Fig. 1 correlated with the fluids contents associated to the deep hydrothermal plume and the shallow gas releasing evidence.

184


represents the final conduit of this complex shallow circulation system, characterized by the joint interaction of the buried fault and the cap rock, whose roles tread along a fine balance of pressure controlled by the deeper magmatic system and the shallow structures. Acknowledgements This work has been a part of the MED-SUV project. MED-SUV has received funding from the European Union's Seventh Program for research, technological development and demonstration under the grant agreement No 308665. ReflexW© by K.J. Sandmeier, Karlsruhe and Move2017.1 academic license for University di Napoli Federico II, were used for some processing, interpretation step and figures. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.jvolgeores.2018.03.025. References Aiuppa, A., Tamburello, G., Napoli, R., Cardellini, C., Chiodini, G., Giudice, G., Grassa, F., Pedone, M., 2013. First observations of the fumarolic gas output from a restless caldera: implications for the current period of unrest (2005–2013) at Campi Flegrei. Geochem. Geophys. Geosyst. 14 (10), 4153–4169. Akaike, H., 1974. A new look at the statistical model identification. IEEE Trans. Autom. Control (6), 716–723. Amoroso, O., Ascione, A., Mazzoli, S., Virieux, J., Zollo, A., 2014. Seismic 796 imaging of a fluid storage in the actively extending Apennine mountain belt, southern Italy. Geophys. Res. Lett. 41:3802–3809. https://doi.org/10.1002/2014GL060070. Amoroso, O., Russo, G., De Landro, G., Zollo, A., Garambois, S., Mazzoli, S., Parente, M., Virieux, J., 2017a. From Velocity and Attenuation Tomography to Rock Physical Modeling: inferences on fluid-driven earthquake processes at the Irpinia fault system in Southern Italy. Geophys. Res. Lett. 44. https://doi.org/10.1002/2016GL072346. Amoroso, O., Festa, G., Bruno, P.P., D'Auria, L., De Landro, G., Di Fiore, V., Gammaldi, S., Maraio, S., Pilz, M., Roux, P., Russo, G., Serlenga, V., Serra, M., Woith, H., Zollo, A., 2017b. Integrated tomographic methods for seismic imaging and monitoring of volcanic caldera structures and geothermal areas. J. Appl. Geophys. https://doi.org/ 10.1016/j.jappgeo.2017.11.012 (ISSN 0926-9851). Benz, H., et al., 1996. Three-dimensional P and S-wave velocity structure of Redoubt Volcano, Alaska. J. Geophys. Res. 101 (4), 8111–8128. Bruno, P.G., Maraio, S., Festa, G., 2017. The shallow structure of Solfatara Volcano, Italy, revealed by dense, wide-aperture seismic profiling. Sci. Rep. 7, 17386. https://doi.org/ 10.1038/s41598-017-17589-3. Byrdina, S., Vandemeulebrouck, J., Cardellini, C., Legaz, A., Camerlynck, C., Chiodini, G., Lebourg, T., Gresse, M., Bascou, P., Motos, G., Carrier, A., Caliro, S., 2014. Relations between electrical resistivity, carbon dioxide flux, and self-potential in the shallow hydrothermal system of Solfatara (Phlegrean Fields, Italy). J. Volcanol. Geotherm. Res. 283:172–182. https://doi.org/10.1016/j.jvolgeores.2014.07.010. Caliro, S., Chiodini, G., Moretti, R., Avino, R., Granieri, D., Russo, M., Fiebig, J., 2007. The origin of the fumaroles of La Solfatara (Campi Flegrei, south Italy). Geochim. Cosmochim. Acta 71 (12), 3040–3055. Chiao, L.-Y., Kuo, B.-Y., 2001. Multiscale seismic tomography. Geophys. J. Int. 145, 517–527. Chiarabba, C., Amato, A., Boschi, E., Barberi, F., 2000. Recent seismicity and tomographic modeling of the Mount Etna plumbing system. J. Geophys. Res. 105 (5), 10923–10938. Chiarabba, C., Piccinini, D., De Gori, P., 2009. Velocity and attenuation tomography of the Umbria Marche 1997 fault system: evidence of a fluid-governed seismic sequence RID G-4780-2011. Tectonophysics 476:73–84. https://doi.org/10.1016/j. tecto.2009.04.004 (ISSN: 0040-1951). Chiodini, G., Frondini, F., Cardellini, C., Granieri, D., Marini, L., Ventura, G., 2001. CO2 degassing and energy release at Solfatara volcano, Campi Flegrei, Italy. J. Geophys. Res. Solid Earth 106 (B8), 16213–16221. Chiodini, G., Selva, J., Del Pezzo, E., Marsan, D., De Siena, L., D'Auria, L., Bianco, F., Caliro, S., De Martino, P., Ricciolino, P., Petrillo, Z., 2017. Clues on the origin of post-2000 earthquakes at Campi Flegrei caldera (Italy). Sci. Rep. 7:4472. https://doi.org/10.1038/ s41598-017-04845-9. Convertito, V., Zollo, A., 2011. Assessment of pre-crisis and syn-crisis seismic hazard at Campi Flegrei and Mt. Vesuvius volcanoes, Campania, southern Italy. Bull. Volcanol. 73, 767–783. D’Auria, L., Giudicepietro, F., Aquino, I., Borriello, G., Del Gaudio, C., Lo, Bascio, Martini, M., Ricciardi, G.P., Ricciolino, P., Ricco, C., 2011. Repeated fluid-transfer episodes as a mechanism for the recent dynamics of Campi Flegrei caldera (1989-2010). J. Geophys. Res. 116, B04313. https://doi.org/10.1029/2010JB007837. D’Auria, L., Pepe, S., Castaldo, R., Giudicepietro, F., Macedonio, G., Ricciolino, P., Tizzani, P., Casu, F., Lanari, R., Manzo, M., Martini, M., Sansosti, E., Zinno, I., 2015. Magma injection beneath the urban area of Naples: A new mechanism for the 2012–2013 volca nic unrest at Campi Flegrei caldera. Sci. Rep. 5 (13100), 1–11.

De Landro, G., Serlenga, V., Russo, G., Amoroso, O., Festa, G., Bruno, P.P., Gresse, M., Vandemeulebrouck, J., Zollo, A., 2017. 3D ultra-high resolution seismic imaging of shallow Solfatara crater in Campi Flegrei (Italy): new insights on deep hydrothermal fluid circulation processes. Sci. Rep. 7 (1), 3412. https://doi.org/10.1038/s41598-01703604-0. De Siena, L., Del Pezzo, E., Bianco F., 2010. Campi Flegrei seismic attenuation image: evidences of gas resevoirs, hydrotermal basins and feeding systems. J. Geophys. Res. 115 (B0), 9312–9329. Di Stefano, R., Chiarabba, C., Chiaraluce, L., Cocco, M., De Gori, P., Piccinini, D., Valoroso, L., 2011. Fault zone properties affecting the rupture evolution of the 2009 (Mw 6.1) L'Aquila earthquake (central Italy): insights from seismic tomography. Geophys. Res. Lett. 38 (L10310). Di Vito, M.A., Isaia, R., Orsi, G., Southon, J., De Vita, S., d'Antonio, M., Pappalardo, L., Piochi, M., 1999. Volcanism and deformation since 12,000 years at the Campi Flegrei caldera (Italy). J. Volcanol. Geotherm. Res. 91 (2), 221–246. Festa, G., et al., 2015. RICEN: Repeated InduCed Earthquakes and Noise @Solfatara, Campi Flegrei. AGU Fall Meeting, San Francisco, pp. 14–18. Gammaldi, S., 2018. High-resolution P-wave Imaging from Active Seismic Data in the Solfatara Crater. University of Naples Federico II, Italy (Ph.D. Thesis in Structural, Geotechnical Engineering and Seismic Risk). Gibson, B., Larner, K., 1984. Predictive deconvolution and the zero-phase source. Geophysics 49, 379–397. Goldberg, X., 1989. Genetic Algorithm in Search, Optimization and Machine Learning. Addison-Wesley Pub Co., Boston. Gresse, M., Vandemeulebrouck, J., Byrdina, S., Chiodini, G., Revil, A., Johnson, T.C., ... Metral, L., 2017. Three-dimensional electrical resistivity tomography of the Solfatara crater (Italy): implication for the multiphase flow structure of the shallow hydrothermal system. J. Geophys. Res. Solid Earth 122:8749–8768. https://doi.org/10.1002/ 2017JB014389. Gunasekera, R.C., Julian, B.R., Foulger, G.R., 2003. Reservoir depletion at The Geysers geothermal area, California, shown by four-dimensional seismic tomography. J. Geophys. Res. 108 (B3):2134. https://doi.org/10.1029/2001JB000638. Isaia, R., Marianelli, P., Sbrana, A., 2009. Caldera unrest prior to intense volcanism in Campi Flegrei (Italy) at 4.0 ka B.P.: implications for caldera dynamics and future eruptive scenarios. Geophys. Res. Lett. 36, L21303. https://doi.org/10.1029/ 2009GL040513. Isaia, R., Vitale, S., Di Giuseppe, M.G., Iannuzzi, E., Tramparulo, F.D.A., Troiano, A., 2015. Stratigraphy, structure and volcano-tectonic evolution of Solfatara maar-diatreme (Campi Flegrei, Italy). Geol. Soc. Am. Bull. 1485-1504 (9), 127. Kilburn, C.R.J., De Natale, G., Carlino, S., 2017. Progressive approach to eruption at Campi Flegrei caldera in southern Italy. Nat. Commun. 8, 15312. https://doi.org/10.1038/ ncomms15312. Laigle, M., Hirn, A., 1999. Explosion-seismic tomography of amagmatic body beneath Mount Etna; volatile discharge and tectonic control of volcanism. Geophys. Res. Lett. 26 (17), 2665–2668. Lees, J.M., 1992. The magma system of Mount St. Helens: non-linear high resolution Pwave tomography. J. Volcanol. Geotherm. Res. 53 (1–4), 103–116. Lumley, D.E., 2001. Time-lapse seismic reservoir monitoring. Geophysics 66:50–53. https://doi.org/10.1190/1.1444921. Okubo, P.G., Benz, H.M., Chouet, B.A., 1997. Imaging the crustal magma sources beneath Mauna Loa and Kilauea volcanoes, Hawaii. Geology 25 (10), 867–870. Orsi, G., de Vita, S., Di Vito, M., 1996. The restless, resurgent Campi Flegrei nested caldera (Italy): constraints on its evolution and configuration. J. Volcanol. Geotherm. Res. 74, 179–214. Orsi, G., Di Vito, M.A., Isaia, R., 2004. Volcanic hazard assessment at the restless Campi Flegrei caldera. Bull. Volcanol. 66:514–530. https://doi.org/10.1007/s00445-0030336-4. Pasquet, S., Holbrook, W.S., Carr, B.J., Sims, K.W.W., 2016. Geophysical imaging of shallow degassing in a Yellowstone hydrothermal system. Geophys. Res. Lett. 43 (23):2016. https://doi.org/10.1002/2016GL071306. Patanè, D., Barberi, G., Cocina, O., De Gori, P., Chiarabba, C., 2006. Time-resolved seismic tomography detects magma intrusions at Mount Etna. Science 313 (5788): 821–823. https://doi.org/10.1126/science.112772. Petrosino, S., Damiano, N., Cusano, P., Di Vito, M.A., de Vita, S., Del Pezzo, E., 2012. Subsurface structure of the Solfatara volcano (Campi Flegrei caldera, Italy) as deduced from joint seismic-noise array, volcanological and morphostructural analysis. Geochem. Geophys. Geosyst. 13:Q07006. https://doi.org/10.1029/ 2011GC004030. Press, W.H., Flannery, B.P., Teukolsky, S.A., Vetterling, W.T., 1986. Numerical Recipes. Cambridge University Press, Cambridge. Rawlinson, N., Hauser, J., Sambridge, M., 2007. Seismic ray tracing and wavefront tracking in laterally heterogeneous media. Adv. Geophys. 49, 203–267. Robinson, E.A., Saggaf, M., 2001. Klauder wavelet removal before vibroseis deconvolution. Geophys. Prospect. 49, 335–340. Serlenga, V., de Lorenzo, S., Russo, G., Amoroso, O., Garambois, S., Virieux, J., Zollo, A., 2016. A three-dimensional QP imaging of the shallowest subsurface of Campi Flegrei offshore caldera, southern Italy. Geophys. Res. Lett. 43 (21):11,209–11,218. https:// doi.org/10.1002/2016GL071140. Serra, M., Festa, G., Roux, P., Gresse, M., Vandemeulebrouck, J., Zollo, A., 2016. A strongly heterogeneous hydrothermal area imaged by surfacewaves: the case of Solfatara, Campi Flegrei, Italy. Geophys. J. Int. 205:1813–1822. https://doi.org/10.1093/gji/ ggw119 (April 07). Villasenor, A., et al., 1998. Three-dimensional P wave velocity structure of Mt. Etna, Italy. Geophys. Res. Lett. 25 (11), 1975–1978. Virieux, J., 1991. Fast and accurate ray tracing by Hamiltonian perturbation. J. Geophys. Res. 96, 579–594.

S. Gammaldi et al. / Journal of Volcanology and Geothermal Research 357 (2018) 177–185 Vitale, S., Isaia, R., 2014. Fractures and faults in volcanic rocks (Campi Flegrei, southern Italy): insight into volcano-tectonic processes. Int. J. Earth Sci. 103:801–819. https:// doi.org/10.1007/s00531-013-0979-0. Whitley, D., 1994. A Genetic Algorithm Tutorial. Samizdat Press http://samizdat.mines. edu/ga tutorial. Zollo, A., D'Auria, L., De Matteis, R., Herrero, A., Virieux, J., Gasparini, P., 2002. Bayesian estimation of 2-D P-velocity models from active seismic arrival time data: imaging of

185

the shallow structure of Mt Vesuvius (Southern Italy). Geophys. J. Int. 151, 566–582, 2002. Zollo, Z., Maercklin, N., Vassallo, M., Dello Iacono, D., Virieux, J., Gasparini, P., 2008. Seismic reflections reveal a massive melt layer feeding Campi Flegrei caldera. Geophys. Res. Lett. 35, L12306. https://doi.org/10.1029/2008GL034242.