Martinelli, Francesco

Active volcanoes are complex, poorly predictable systems that can pose a threat to humans and their infrastructures. As such, it is important to improve as much as possible the understanding of their behavior. The Stromboli volcano, in Italy, is one of the most active volcanoes in the world, and its almost persistent activity is documented since centuries. The persistent background activity is sometimes interrupted by much more energetic, dangerous episodes. The Istituto Nazionale di Geofisica e Vulcanologia (Italy) set up the interdisciplinary “UNO” project, aimed to understand when the Stromboli volcano is about to switch from the ordinary to the extraordinary activity. The UNO project includes an outstanding variety of research activities, such as sampling in the field, the modeling of Stromboli topography from ALS technique and satellite data, the 3D numerical simulations of ballistic trajectories, or the set up of an ultrasonic microphones system. Key to the success of the project is the collection of integrated high spatial and temporal resolution data and their joint analyses in a shared relational database. We present here the simplified logical model of such database, focusing on the identification of entities and their relationships.

We describe the main structure and outcomes of the new probabilistic seismic hazard model for Italy, MPS19 [Modello di Pericolosità Sismica, 2019]. Besides to outline the probabilistic framework adopted, the multitude of new data that have been made available after the preparation of the previous MPS04, and the set of earthquake rate and ground motion models used, we give particular emphasis to the main novelties of the modeling and the MPS19 outcomes. Specifically, we (i) introduce a novel approach to estimate and to visualize the epistemic uncertainty over the whole country; (ii) assign weights to each model components (earthquake rate and ground motion models) according to a quantitative testing phase and structured experts’ elicitation sessions; (iii) test (retrospectively) the MPS19 outcomes with the horizontal peak ground acceleration observed in the last decades, and the macroseismic intensities of the last centuries; (iv) introduce a pioneering approach to build MPS19_cluster, which accounts for the effect of earthquakes that have been removed by declustering. Finally, to make the interpretation of MPS19 outcomes easier for a wide range of possible stakeholders, we represent the final result also in terms of probability to exceed 0.15 g in 50 years.

Horizon 2020 Project ID: 731070 Duration: From 2018-02-01 to 2021-11-30 Funded under: H2020-EU.1.4.1.2. – Integrating and opening existing national and regional research infrastructures of European interest EUROVOLC is Coordinated by the Icelandic Meteorological Office. Project Coordinator: Dr. Kristín S. Vogfjörd.

EUROVOLC Virtual Accesses offer the opportunity to anyone with a web access to use online tools related to volcanological research. The Volcano Dynamics Computational Center at INGV in Pisa offers the access to a suite of fast-performing numerical codes aimed at modeling different aspects of volcano dynamics: solwcad: Fortran code that computes the fully non-ideal, multi-component, compositional-dependent saturation surface of H2O+CO2 in silicate melts over P-T-composition conditions relevant to magmatism and volcanism. Calculations allow to either 1) determine the partition of H2O and CO2 between the melt and gas phase, or 2) determine the entrapment pressure and corresponding gas phase composition from dissolved amounts; MAMMA: FORTRAN90 code designed to solve a conservative model for magma ascent in a volcanic conduit, described as a compressible multi-component two-phase flow. The system of conservation equations considers the effects of the main processes that magmas experience during ascent, such as crystallization, rheological changes, fragmentation, physical interaction with conduit walls, out-gassing and degassing. The model is capable of describing conduits with elliptical cross sections and depth-dependent dimensions; PyBOX: Python/Fortran90 code that solves the so-called “box model” equations describing the kinematics of a pyroclastic density current over a flat surface and in a steady atmosphere. The model integrates a procedure to account for blockage of PDCs by a rugged topography imported as a ASCII file, by adopting the so-called “energy-conoid” approach. Virtual Access will include an interface to import the DEM file and input parameters and to visualize georeferenced maps of invasion and plots of decaying dynamic pressure.

During the process for assessing seismic hazard there are some recurring operations which can be automated, which are about analysing the input data elements. The recurring operations are based on standard procedures and well known algorithms, and are performed a number of times in the range of tens, or even hundreds. It is therefore important to guarantee the correctness of the elaborations, and to reduce the human intervention which is always a potential source of errors. Specifically, we identified the following operations to be automated: the declustering of a catalog, the estimate of the catalogue completeness, the estimation of seismicity rates of source areas according to the Gutenberg-Richter distribution (G-R; [Gutenberg & Richter, 1944]). To perform the above operations, we developed a web application, AutoRate. The application allows to use databases (including geographical databases) shared between the users, to have the code always updated and available to all the users, to use shared dedicated resources to perform the elaborations removing the load on personal computers. The application offers the users to choose different approaches for each operation; additionally, it is explicitly designed to host new approaches. The output format of each type of operation is independent from the selected algorithm. Each output format is designed to allow a rapid and simple comparison and to be immediately usable as input for the computation of the seismic hazard.

In 2015, the Seismic Hazard Centre (CPS) of the Istituto Nazionale di Geofisica e Vulcanologia (INGV) was commissioned to engage and coordinate the national community with the aim of elaborating a new reference seismic hazard model, which is expected to be released in mid 2018. CPS outlined a roadmap to describe the main features of this complex endeavour, including the different scientific tasks, milestones and timelines. The scientific tasks focus their work on i) improving the quality and the accuracy of the input data (e.g. historical seismic catalogue, seismotectonic zonation, etc.); ii) building new earthquake rate models based on these new input data, iii) selecting the most proper ground motion prediction equations, iv) testing the overall seismic hazard model as well as each component; v) combining the results of the statistical testing phase and the outcome of an expert's elicitation session to assign a weight to each component of the final seismic hazard model. The new seismic hazard model is based on an innovative coherent probabilistic framework, which allows a proper description of the aleatory variability and epistemic uncertainty, and the validation of the seismic hazard model. Here, we describe the progresses made up to now, the comparison between the new and the official national model, and finally we discuss the scientific aspects that have the most significant impact on the new picture of PSHA in Italy.

The Italian reference seismic hazard model was released in 2004, but it has been adopted for the definition of seismic zones in 2006 and for building code only in 2009. At the beginning of 2015 the Seismic Hazard Center (CPS) of INGV was commissioned to coordinate the national scientific community with the aim of elaborating a new reference seismic hazard model, mainly finalized to the update of seismic code. The CPS designed a roadmap to release within 2 years a significantly renewed model, with regard both to the updated input elements and to the strategies to follow, in order to obtain a shared and largely accepted PSHA. The main requirements of the model were discussed in meetings with the experts on earthquake engineering. A public call was opened according to a transparent procedure; we received 24 proposals from many national institutions. The activities were organized in 6 tasks: project coordination, input data, seismicity models, ground motion prediction equations, computation and rendering, validation. In the first phase, the working groups of each task worked separately; in the second phase of the project they collaborated to release a final model. During the project, many scientific aspects were carefully considered, as in many other seismic hazard projects: the use of a declustered catalogue versus a non declustered one, the adoption of the logic-tree approach instead of an ensemble modeling, the definition of objective strategies to assign the weight to each single model, and so on.

In the frame of the Italian research project INGV-DPC S2 (http://nuovoprogettoesse2.stru.polimi.it/), funded by the Dipartimento della Protezione Civile (DPC; National Civil Protection Department) within the agreement 2007-2009, a tool for probabilistic seismic hazard assessment (PSHA) was developed. The main goal of the project was to provide a flexible computational tool for PSHA; the requirements considered essential for the success of the project included: • ability to handle both stationary and non-stationary earthquake time-occurrence models; • ability to use ground-motion prediction models that are not parametric equations but probabilistic "footprints" of the intensities generated by earthquakes of known magnitude and focal characteristics. Usually, these footprints are results of ground motion simulations. Some commonly used programs (e.g., FRISK, by McGuire, 1978; SEISRISK III, by Bender and Perkins, 1987) and more recent and state-of-the-art tools (e.g. OpenSHA, by Field et al., 2003, http://www.opensha.org; OpenQuake, http://openquake.org) for PSHA were analyzed. It was decided to focus on CRISIS2007, which was already a mature and well known application (e.g., Kalyan Kumar and Dodagoudar, 2011; Teraphan et al., 2011; D’Amico et al., 2012; see also http://ecapra.org/CRISIS-2007), but also suitable for additional development and evolution since its source code is freely available on request. The computational tool resulted in an extensive redesign and renovation of the previous CRISIS2007 version. CRISIS is a computer program for PSHA, originally developed in the late 1980's using Fortran as programming language (Ordaz, 1991). In this format, still without a graphical user interface (GUI), it was distributed as part of SEISAN tools (Ottemöller et al., 2011). Ten years later, a GUI was constructed, generating what was called CRISIS99 (Ordaz, 1999). In this version, all the graphic features were written in Visual Basic, but the computation engine remained a Fortran dynamic link library. The reason for the use of mixed-language programming was that computations in Visual Basic were extremely slow. Around 2007 the program was upgraded, in view of the advantages offered by the object-oriented technologies. An object-oriented programming language was required and the natural choice was Visual Basic.Net. In the new version (called CRISIS2007), both the GUI and the computation engine were written in the same language. Finally, in the frame of the mentioned S2 project, starting from 2008, the program was split into two logical layers: core (CRISIS Core Library) and presentation (CRISIS2008). In addition, a new presentation layer was developed for accessing the same functionalities via Web (CRISISWeb). It is worth noting that CRISIS has been mainly written by people that are, at the same time, PSHA practitioners. Therefore, the development loop has been relatively short, and most of the modifications and improvements have been made to satisfy the needs of the developers themselves.

A probabilistic seismic hazard analysis (PSHA) was carried out for the SE sector of Sicily, an area characterized by highest levels of seismic hazard in Italy and high exposure, both in terms of cultural heritage and of critical industrial facilities. Compared to the Italian reference PSH map (MPS04), this study is based on most updated information about regional seismic sources and ground-motion attenuation. Epistemic uncertainties associated with the input elements of the computational model were taken into account following a logic-tree approach. Special care was devoted to define the regional source zones model by considering four alternative models, which share the zones defining the boundary conditions of the study area but differ in the seismotectonic characterization of SE Sicily. Seismic hazard was assessed in terms of PGA, PGV, acceleration and displacement elastic response spectra on rock for four return periods (30, 50, 475, 975 years). A disaggregation analysis was then performed for some sites of interest. Results confirm the very high hazard of the area, with expected values of PGA (at 10% probability of exceedance in 50 years) slightly higher than the reference MPS04 map. Strong differences emerge instead between the acceleration response spectra of this study and the reference ones, for the longest return periods.