Tilt vectors at the three tiltmeters CMP (blue), ECO (green) and HDM (red).The curves are the cumulative tilt variation (hodograph) recorded in 2021. The black dots superimposed mark the monthly tilt. The tilt increase on the NS component is northward down with respect to the site, the tilt increase on the EW component is eastward down. Grid spacing tilt variation of 10 μradians.

Campi Flegrei ground velocity maps retrieved by InSAR analysis from several SAR missions. [A] ERS (1992-2002), [B] Envisat (2003-2010) and [C] Cosmo-SkyMed (2011-2020) SAR images.

Objectives

WP 4 aims at addressing the long-term and short-term deformation styles at Campi Flegrei, considering most of the available geodetic data and analyzing new data. The observed deformation is interpreted by inverse modelling and by building advanced models of the plumbing system. The recent activity of uplift/deflation, detected by several modern geodetic techniques, will be considered and related to the study of the historical deformations at Campi Flegrei. Also, a new dataset of archeological elevation changes of Campi Flegrei is analyzed and modelled, disclosing many details on the past activity at Campi Flegrei. The WP is organized with two tasks, task 4.1 that collects the geodetic and gravimetric data, with task 4.2 mostly devoted to data interpretation and modelling.

Task 4.1 Geodetic and gravimetric data analysis

An unprecedented set of geodetic data at Campi Flegrei from leveling, tilt, InSAR (Interferometric Synthetic Aperture Radar) techniques, and seafloor measurements will be analyzed and collected.

In particular

• the long-term archive of leveling data along the Pozzuoli coast is integrated with geological and archaeological observations. Archeological archive data since 35 b.C. is analyzed for the first time.
• The database is integrated with data from the tiltmeter network, presently including 10 stations, available since 1998.
• The capability of InSAR data is fully exploited in the project to create displacement maps and time series from the whole C-band Envisat data archive (2002-2010) and Sentinel data (since 2014) from the European Space Agency, along with ASI’s (Agenzia Spaziale Italiana) X-band Cosmo-SkyMed satellite data since 2010.
• The seafloor measurements produced by the marine infrastructure MEDUSA are included in the WP, and given the large submerged portion of the CF caldera, these data will significantly improve the knowledge of the ground deformations
• Possibly, archive gravimetric data will be taken into account to characterize the hydrothermal/magmatic nature of the past and present unrests.

Task 4.2 Long-term and short term deformation history and characterization of eruptive and unrest volcanic sources

The objective of this task is analyzing short-term and long-term deformation styles, and their relationship with the volcanic activity and plumbing system at Campi Flegrei. Inverse modelling is performed to interpret the available geodetic datases to retrieve the deformation system at Campi Flegrei. The large amount of data available is considered with different time windows. In particular, one objective is to study the past behaviour of Campi Flegrei by modelling the archeological data since 35 BC. Also, modern geodetic data from the local networks are modelled with a similar approach to understand the volcanic activity through centuries at Campi Flegrei. Numerical models of the volcanic source and plumbing system are developed based on the updated knowledge of the volcanic activity and including realistic features of the Campi Flegrei volcanic system. The models account for new insights derived from volcanological/petrological (WP1), geochemical (WP2) and seismological (WP3) studies. The causes for the subsidence/uplift and gravimetric variations are addressed considering the effect of several factors which may act along with magmatic activity at depth, such as - for instance - temperature and pressure variations within the hydrothermal system, presence of elastic discontinuities and inelastic rheology, and the action of structural discontinuities such as the faults related to both caldera collapse and resurgence of its floor. The models are applied both to the current unrest of the Campi Flegrei, and to characterize the past long-term pre-eruptive phases. In collaboration with task 1.2, initial stress scenarios will be produced for dyke propagation simulations.

Reconstruction of the elevation (meters above the sea level in 1905 – first leveling by Istituto Geografico Militare) through time within the Campi Flegrei caldera, obtained integrating geological, historical and archaeological data.

Scheme of the Volcanic and Seismic source Modelling (VSM) tool used to perform geodetic data inversions. The input of VSM consists of at least one geodetic dataset. The forward model is one or more analytical source(s). One of the two available inversion algorithms is selected. After data inversion is completed, several products are generated as output.

Reconstruction of elevation changes since 35 BP

The Reconstruction of the ground movements is performed with an unprecedented integrated analysis of geomorphological, sedimentological, paleontological, archaeological and historical data. This procedure allowed a detailed and quantitative reconstruction of the evolution of the ground displacements since 35 BP, evidencing the deformations related to the 1538 Mt. Nuovo eruption along the coastline of the Pozzuoli Bay.

Geodetic Data Analysis

InSAR analysis applied to the long dataset of ERS, Envisat and Cosmo-SkyMed images acquired  from 1992 to 2020 was performed by multi-temporal Interferometric Point Target Analysis (IPTA) approach. Such approach allows to select the interferograms to be used to estimate both ground deformation rate and displacement time series by maxima baseline criteria. The chosen interferometric pairs are then estimated, filtered by Goldstein filtering and unwrapped by minimum cost flow algorithm. Hence, we select the InSAR point targets by accepting the points with temporal coherence value equal or greater than 0.4 and the final solution is found by Singular Value Decomposition analysis.

The outcomes show a deflation phase of Campi Flegrei imaged by ERS data from 1992 to 2002, followed by a first slow inflation phase until 2010 detected by Envisat data and then the fast inflation with uplift values up to 10 cm/yr started in 2011 as detected by Cosmo-SkyMed data.

Geodetic data modelling

The large amount of geodetic data collected is analyzed by an inverse modelling scheme, to retrieve the active deformation sources at Campi Flegrei. Data inversion is performed using the Volcanic and Seismic source Modelling (VSM) tool. VSM is an open source Python tool that accepts input data from most of the geodetic techniques (InSAR, GPS, levelling, EDM, tilt and strain data), alone or in combination. The VSM tool allows the user to choose one or more geometrical source as forward model among sphere, spheroid, ellipsoid,fault, and sill. Two inversion tools are available, one is a global optimization algorithm based on the Voronoi cells and the second follows a probabilistic approach to parameters estimation based on the Bayes theorem. VSM can be executed as a Python script, in a Jupyter Notebook environment or by its Graphical User Interface. VSM, developed at INGV, supports all the activities related to geodetic data inverse modelling in the project.

Location of the earthquakes of the swarm of 7/9/2012 using the NLLoc program with a 3D velocity model.. Circle sizes are proportional to the earthquake magnitude. The beach balls are obtained using the program FPFit. The numbers within the beach balls indicate the minutes and seconds of the earthquake occurrence.

Position of the seismic array deployed during the temporary experiment, and position of CPIS permanent seismic station. Arrows indicate the direction of provenance of the signal derived from FK and polarization analysis. Red star indicates the position of the Pisciarelli boiling pool; b) back-azimuth and c) seismic amplitude associated with seismic array detections indicate a stationary and fixed seismic source. d) Seismic amplitude measured at CPIS permanent seismic stations show an extreme coherence with detection amplitude derived from array analysis.

Task 3.1:

We reconstructed the history of the seismic network of Osservatorio Vesuviano since 2012 to evidence any station variation that could give anomalies in the cross correlation analysis. We organized the data set and configured a work station where to allocate the analysed data.  We computed the cross correlation on the vertical component of the seismic records within different frequency bands. For the analysis we selected 20 stations of the permanent seismic network of Osservatorio Vesuviano. Not all of them have been working since 2012. We performed the cross correlations for all the couple of stations in the overlapping working periods. The correlations have been stored to be used for the tomographic analysis. Some correlations show picks that needs to be better analysed. We are going to analyse the cross correlation functions in terms of path characteristics.

Task 3.2:

We statistically analyzed the seismic catalogue of Campi Flegrei from 2000 to 2020 to understand the temporal variations of seismic parameters and correlated them with the trend of the geochemical parameters estimated from the fumarolic gases recorded at Solfatara and with the deformations of the central part of the caldera. A detailed analysis of the Mc of the whole catalogue, with a particular attention to its variations with time and space, has been. The analysis of the catalogue was accompanied by the simultaneous computation of the theoretical sensitivity analysis of the seismic network to corroborate the empirical results and to have an Mc estimate also for the areas of the caldera where the seismicity was too low to allow an empirical estimate. From August 2000 to April 2020 about 1700 earthquakes have been recorded. Most of them are located in the Solfatara/Pisciarelli area and we estimated a b value of 0.92 ± 0.03 for the whole period.  A decrease in the average inter-distance and in the average depth is evident since the end of 2015, around October 7, when a swarm occurred at the end of a period of uplift increment that stopped in coincidence with the swarm itself.  All the analyses are well described in Tramelli et al. (2021) where the authors led back to a common cause for all the temporal variation of the observable taken into consideration. The fluid heating or its pressurization, clearly revealed by the CO/CO2 increase in time, can indeed cause all the monitored anomalies. Injection of magmatic fluids within the hydrothermal system which feeds the Solfatara/Pisciarelli fumaroles would heat up the system increasing the pore pressure and facilitating the earthquakes occurrence being the medium less prone to bear stress and being the faults lubricated.

In addition, we analysed in detail six main swarms recorded in Campi Flegrei occurred in 2000, 2012, 2015, 2018, 2019 and 2020 in order to contribute to the understanding of the processes involved in the long term unrest of this densely populated caldera. We re-located the earthquakes and estimated the focal mechanisms of the higher-magnitude ones within these swarms. We compared the locations and focal mechanisms with the fault systems of the caldera and with the tomographic images and we found a good agreement between them. The focal mechanisms are in agreement with the tensional stress induced by the uplift of the central part of the caldera.

Task 3.3:

We conducted a temporary seismic array measurements at Pisciarelli fumarolic field in order to investigate the seismic source process which generate seismic tremor and its relationship with fumarolic outgassing. We deployed a small aperture 4-element 3D seismo-acoustic array equipped with Lennartz 3D-1s seismometers at a distance of ~40m from the fumarolic field. We characterized the seismic wavefield produced by Pisciarelli fumarolic field in terms of back-azimuth, slowness and associated amplitudes. Array data processing on seismic signals allowed to define the coherent data of the measured ground motion from a fixed or a moving source, and, in addition, allows to define the direction of provenance of the seismic wavefield through time and space. The recorded seismc wavefield observed at Pisciarelli can be traced back to a persistent and fixed source located at ~200°N from the array, which is consistent to the position of the fumarolic field. The comparison of the array measurements with the records of the permanent seismic station of the network of Osservatorio Vesuviano CPIS, used as the reference station for fumarolic tremor monitoring, shows an extreme coherence and similarity of the mutual amplitude fluctuations (Giudicepietro et al., 2019). Array analysis also allowed to derive the “apparent velocity” parameter (the inverse of slowness) associated with incident seismic wavefield which gives information on the depth position of the radiating source. The seismic wavefield recorded at Pisciarelli is associated with an apparent velocity of 300-400 m/s, which indicates a very shallow position of the source propagating mainly on the horizontal plane.

Plot of the earthquake magnitudes with time, top, and histogram of the earthquakes per year, bottom.

Seismicity at Campi Flegrei since 2000 and updated to 2020. Left column: location of the earthquakes recorded in the Campi Flegrei since 2000. Right column: hypocenter densities on a 200-m-spaced grid cells. The minimum value is two earthquakes per cell; the maximum is 10 earthquakes per cell. The red arrow indicates the position of the Pisciarelli hydrothermal area. The plots have a vertical exaggeration of 2.

In this WP, we will focus on the analysis of the seismic signals recorded in the Campi Flegrei by INGV-OV seismic network from 2012 to 2019 to highlight any eventual variation and explore its meaning. We plan to analyze the seismic noise recorded in the caldera focusing on its characteristics and variations in time (task 3.1); to statistically analyze the seismic catalogue with the aim to evidence the seismicity characteristics in time and space (task 3.2) and to analyze the seismic fumarolic tremor recorded close to Pisciarelli hydrothermal area to delineate its source mechanism and to verify its possible meaning as indicator of unrest. Finally, we will interpret the achieved results with geochemical, and geodetic data and the OV seismic catalogue to define the best conceptual model supporting all gathered observations.

Task 3.1) Noise-cross correlation for seismic velocity variations.

In this task we plan to apply the noise cross correlation techniques to seismic data recorded by INGV-OV seismic network from 2012 to 2019 at the Campi Flegrei caldera to evidence any velocity changes of the medium with time and to eventually associate them to other geodetical, geochemical and seismological changes.

Despite the seismic studies performed in the last decades using the data of the last bradiseismic crises and/or active experiments executed in 2001 (Zollo et al., 2008), a high detailed image of the shallow portion of the caldera is actually not available and, anyway, not updated. Particularly, noise correlation techniques applied at Campi Flegrei between 2011 and 2015 allowed to identify little velocity changes of the rocky substrate possibly linked to changes in magma/fluid dynamic, as also evidenced by coherent changes of other geochemical or geodetic parameters (Zaccarelli and Bianco, 2017). Part of these data, collected between 2011 and 2013, have been also used to reconstruct the shear wave velocity model but without

analyzing any temporal variation of them (De Siena et al., 2018).

Task 3.2) Statistical analyses of seismic cataloge.

In this task we will perform a statistical analysis of the seismic catalogue of Campi Flegrei. In particular, we will analyze changes with time and space of recurrence rate of local earthquakes.  INGV-OV seismic catalogue shows, indeed, an evident increase of seismic rate from 2016 until today passing from ~ 300 to ~ 800 events per year. Moreover, using the same catalogue, a detailed analysis of the most energetic swarms recorded since 2000 will be performed to help in identifying the faults that are more productive and to correlate them to the outgassing dynamics. The identification of doublets within the earthquakes recorded since 2000 would eventually allow to evidence long term variations in the medium properties, in parallel to the analysis performed with the noise cross correlation (task 3.1).

Task 3.3) Investigating fumarolic tremor source dynamics at Campi Flegrei caldera.

In this task we propose to investigate fumarolic tremor at Campi Flegrei Caldera using data from INGV permanent seismic network and using ad-hoc temporary deployments of seismic arrays. Campi Flegrei Caldera is, indeed, characterized by a high-frequency tremor source generated by fumarole outgassing. During 2012 and 2013 Campi Flegrei caldera experienced a phase of increase in the uplift rate (De Martino et al., 2014, Chiodini et al., 2015; Tamburello et al., 2019), which was followed by increases of shallow earthquake occurrence rate and also by a significant increase of fumarolic tremor amplitude (Giudicepietro et al., 2019). The aim of this task is to characterize tremor source mechanism and its variation through time and space within the last decade in order to improve our understanding of the magma plumbing system and associated degassing dynamics. The results will also improve our ability to detect any volcanic tremor precursors associated with the next eruption, with strong implications on risk management of Neapolitan urbanized area.