• numerical modelling
• seismic hazard
• seismic microzonation
• ambient vibrations
• seismic response analysis

Seismic hazard assessment (i.e., a forecast of seismic ground motion expected in the next tens of years in the Italian territory) represents a main goal of seismological studies since it represents the basic element for planning risk reduction policies by addressing available resources where they are more urgent and where their effectiveness can be maximized. Well-consolidated approaches exists for seismic hazard assessment at regional scale, whose outcomes constitute a general reference for anti-seismic design of structures. On the other hand, recent earthquakes occurred in Italy show that seismic effects show a high degree of lateral variability at the scale of hundreds to thousands of meters (local seismic response) that cannot be captured by regional scale hazard maps. This makes mandatory and urgent the definition of local hazard maps (at the scale of a single municipality) providing a more realistic picture of expected seismic ground motion. This poses a number of methodological problems that are generally ignored when regional scale hazard maps are developed. In fact, in these last maps, large scale propagation effects and geometry of seismic sources play a major role, while the effects of small-scale seismo-stratigraphical and morphological features is ignored by referring to a ‘standard’ configuration (rigid deposits) rarely found in actual situations. On the other hand, the former features are responsible for large modifications of ground motion at local scale and require very detailed survey procedures to be detected and characterized. To be effective, these last studies (Seismic Microzonation) must be performed extensively over wide areas in all municipalities exposed to potentially damaging earthquake. This prevents the application of costly procedures for subsoil exploration and requires the development of cheap and effective approaches allowing a general application of local seismic hazard assessment also where resources available on purpose are scarce. In the last years, on behalf of National Civil Protection and Regional Authorities, the Italian scientific community has been deeply involved in developing and applying such approaches, which have been initially formalized in the Italian Guidelines for Seismic Microzonation or IGSM (WGSM, 2008) providing guidelines for Seismic Microzonation. An important field application of these guidelines after the L’Aquila earthquake (2009) showed effectiveness and limitations of these guidelines and provided a new impulse to these studies. This evolution in presently going on and the studies summarized in this Thesis represent a contribution to improve and validate procedures and geophysical exploration tools on purpose developed for supporting and improving seismic microzoning studies.
The Thesis includes 6 chapters and respective contents reflect number of papers I published during the PhD activity.
After a short introduction (Chapter 1) and a general summary of main problems involved in the extensive application of seismic microzoning tools, in Chapter 2, a methodology is presented for the definition of an effective tool to estimate local amplification factors based on low-cost geophysical prospecting tools. In fact, an extensive application of seismic microzoning studies requires the definition of specific proxies to allow local practitioners to quantify litho-stratigraphical amplification phenomena on the basis of procedures simple enough to allow a widespread application. These tools should be specialized to the litho-stratigraphical configurations representative of the study area. A procedure is here described to provide such a tool based on extensive numerical simulations taking advantage of geological/geotechnical information made available by regional/national Authorities. This procedure is quite general and could be applied in several contexts and, in particular, in developing countries characterized by low seismicity rates or where extensive accelerometric databases are lacking. As an example, an application of the above procedure developed on behalf of the Tuscany Regional Administration (in Central Italy) is presented.
In Chapter 3 an important methodological aspect is addressed, concerning the role of vertical ground motion components in seismic response analysis. This element has been generally discarded in common approaches but revealed to be of great importance during the seismic sequence that stroke the Po Plain in 2012. In fact, in the seismic codes currently applied, vertical components accounted for in anti-seismic design are considered to be a fraction of the horizontal ones and are supposed to play a minor role in the definition of damages expected in existing structures. However, observations carried out during last important seismic events in Italy (L’Aquila in 2009 and Emilia-Romagna in 2012) demonstrate that vertical components were larger than those expected on the basis of current hazard estimates and played an important role in damaging specific structures (masonry buildings, churches, industrial warehouses, etc.) located in the epicentral area. In order to provide a more correct estimate of seismic hazard in the near field, numerical simulations have been carried out to explore features of the vertical ground motion at the top of a sedimentary cover overlying a buried seismogenic sources. Numerical modelling is considered to evaluate the possible effect of vertical ground motion components of input motion on the horizontal seismic response at the surface of a stack of homogenous sedimentary layers. This analysis has been performed at four Italian sites where the local Vs profile was available down to seismic bedrock. Computations show that the effect of vertical components on horizontal ground seismic response is frequency dependent and changes as a function of the local Vs profile and of the accelerometric time series. These outcomes suggest that the common practice of considering only horizontal components of input motion may result ineffective and provides in some cases of possible practical interest underconservative outcomes.
After that in Chapter 4 some case studies are presented, which respectively represent seismic microzoning studies performed on behalf of local and national Authorities at different levels by following the IGSM.
The first example concerns the development of a preliminary Microzonation study (Level I, by following IGSM) developed in the frame of the project ‘Rischio Sismico negli Edifici Monumentali-RiSEM’ (Seismic Risk in Monumental Buildings). In this case, a seismo-geological reference model was developed for the historical centre of the town of San Gimignano (Central Italy) in order to evaluate possible small-scale lateral variation of seismic hazard. To this aim, an on-purpose geological and geomorphological survey was performed along with ambient vibration measurements (both in a single station and array configurations) to characterize seismic response in the study area. Expected seismic amplification effects were quantified by considering a simplified approach developed on purpose and described in Chapter 2. This study provides preliminary information supporting site-specific analyses of the local seismic response and makes it possible to identify most critical situations where eventual seismic retrofitting interventions are more urgent to reduce seismic risk.
A second example concerns a more advanced Microzonation study (Level II by following IGSM) relative to the municipality of Montecatini Terme (Central Italy). Here, a previous Microzonation study was improved by implementing a more detailed geophysical survey as required for the second level of microzonation.
The third case study concerns a more advanced microzoning study (Level III by following IGSM) developed in support of seismic reconstruction in the area struck by the recent disastrous seismic sequence (2016-2107) in Central Italy. In that context, the Department of Physics, Earth and Environmental Sciences of the University of Siena was involved in the activities for the reconstruction project such as geophysical survey, but also for supporting the realization of level I, level II and level III Microzonation. In particular in the municipality of Montegallo (near Mount Vettore), a geological model was realized and the relevant geological sections were used to find reference seismostratigraphical log for each residential area. Each log was then used to estimate a one-dimensional seismic response and then to calculate some amplification factors. It was also used a two-dimensional seismic response analysis to evaluate possible topographic and stratigraphic amplification effects.
The fourth case study concerns an extensive study aiming at evaluating seismic response at accelerometric sites to retrieve ground motion at reference soil conditions by deconvolution analysis. To allow a generalized application to large areas where borehole data are generally lacking or appear inadequate for the seismic characterization for soils down to the reference seismic bedrock, cost-effectiveness of the considered procedures is a main issue. Thus, major efforts have been devoted to optimize available information and exploit fast and cheap surface geophysical prospecting. In particular, geological/geomorphological survey and passive seismic prospecting (both in single and multi-station configurations) were jointly considered to reconstruct seismo-stratigraphical site conditions. This information was then used to feed numerical modelling aiming at computing the local seismic response and performing a deconvolution analysis to reconstruct ground motion at reference soil conditions. Major attention was devoted to evaluate and manage uncertainty involved in the procedure and to quantify its effect on final outcomes. An application of this procedure to a set of sites included in the Italian Accelerometric Network is presented.

In Chapter 5, a step is made in the direction of a more comprehensive estimate of seismic effects by considering the possible impact of ground motion on structures. In fact, the expected soil-structure interaction is very important in evaluating seismic risk, because damages enhance when dynamic properties of buildings induce resonance phenomena at frequencies corresponding to those locally amplified by subsoil configuration. Thus, beyond information about seismic response of the local subsoil, retrieving information about dynamic characteristics of building exposed to future events is of paramount importance to identify most critical situations. To this purpose, passive seismic methods may be of great importance due to the relative easiness of their application in several contexts. In this Thesis, two case studies are presented concerning two historical hamlets, which are famous for their historical buildings (in particular for their towers) and they are both included in the UNESCO World Heritage List. The sites are located in regions characterized by medium level of seismic hazard. Therefore, mitigating earthquake damages is an important goal for conservation of historical buildings. From an economic point of view, this issue is fundamental in order to sustain tourism, which is a relevant source of local economy. A first step in this direction is the evaluation of dynamic response to seismic loads at least in the domain of small strain levels corresponding to the beginning of damage. Several recent studies demonstrate that this task can be achieved efficiently by using suitable single station asynchronous ambient vibration measurements.
The first case study is performed in San Gimignano (Tuscany, Central Italy). In order to improve the knowledge about the seismic risk of this cultural heritage, the Regione Toscana (Tuscany Regional Administration) promoted the RISEM project (RIschio Sismico negli Edifici Monumentali - Seismic Risk in Monumental Buildings), involving the universities of Florence and Siena. A part of this project was devoted to the estimate of possible damages expected at each tower as a consequence of possible future earthquakes. In this frame, ambient-vibration monitoring was used along with measurements inside the buildings to contribute to evaluate dynamical response of the towers. The second case study was the San Marino Historical centre. The same technique described in the previous case has been applied to three medieval towers located in San Marino and allowed identifying fundamental (elastic) resonance frequency, a key parameter for assessing seismic behaviour of historical buildings.