DSpace Collection:

How do volcanic rift zones relate to flank instability? Evidence from collapsing rifts at Etna

Tue, 18 Sep 2012 22:00:00 GMT

Title: How do volcanic rift zones relate to flank instability? Evidence from collapsing rifts at Etna Authors: Ruch, J.; Dipartimento Scienze Geologiche, Università Roma Tre, Roma, Italy; Pepe, S.; National Research Council (CNR), Istituto per il Rilevamento Elettromagnetico dell'Ambiente (IREA), Napoli, Italy; Casu, F.; National Research Council (CNR), Istituto per il Rilevamento Elettromagnetico dell'Ambiente (IREA), Napoli, Italy; Acocella, V.; Dipartimento Scienze Geologiche, Università Roma Tre, Roma, Italy; Neri, M.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Catania, Catania, Italia; Solaro, G.; National Research Council (CNR), Istituto per il Rilevamento Elettromagnetico dell'Ambiente (IREA), Napoli, Italy; Sansosti, E.; National Research Council (CNR), Istituto per il Rilevamento Elettromagnetico dell'Ambiente (IREA), Napoli, Italy Abstract: Volcanic rift zones, characterized by repeated dike emplacements, are expected to delimit the upper portion of unstable flanks at basaltic edifices. We use nearly two decades of InSAR observations excluding wintertime acquisitions, to analyze the relationships between rift zones, dike emplacement and flank instability at Etna. The results highlight a general eastward shift of the volcano summit, including the northeast and south rifts. This steadystate eastward movement (1-2 cm/yr) is interrupted or even reversed during transient dike injections. Detailed analysis of the northeast rift shows that only during phases of dike injection, as in 2002, does the rift transiently becomes the upper border of the unstable flank. The flank's steady-state eastward movement is inferred to result from the interplay between magmatic activity, asymmetric topographic unbuttressing, and east-dipping detachment geometry at its base. This study documents the first evidence of steady-state volcano rift instability interrupted by transient dike injection at basaltic edifices.

A pilot GIS database of active faults of Mt. Etna (Sicily): A tool for integrated hazard evaluation

Mon, 31 Dec 2012 23:00:00 GMT

Title: A pilot GIS database of active faults of Mt. Etna (Sicily): A tool for integrated hazard evaluation Authors: Barreca, G.; Università degli Studi di Catania, Dipartimento di Scienze Biologiche, Geologiche ed Ambientali, Sez. Scienze della Terra, Italy; Bonforte, A.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Catania, Catania, Italia; Neri, M.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Catania, Catania, Italia Abstract: A pilot GIS-based system has been implemented for the assessment and analysis of hazard related to active faults affecting the eastern and southern flanks of Mt. Etna. The system structure was developed in ArcGis® environment and consists of different thematic datasets that include spatially-referenced arc-features and associated database. Arc-type features, georeferenced into WGS84 Ellipsoid UTM zone 33 Projection, represent the five main fault systems that develop in the analysed region. The backbone of the GIS-based system is constituted by the large amount of information which was collected from the literature and then stored and properly geocoded in a digital database. This consists of thirty five alpha-numeric fields which include all fault parameters available from literature such us location, kinematics, landform, slip rate, etc. Although the system has been implemented according to the most common procedures used by GIS developer, the architecture and content of the database represent a pilot backbone for digital storing of fault parameters, providing a powerful tool in modelling hazard related to the active tectonics of Mt. Etna. The database collects, organises and shares all scientific currently available information about the active faults of the volcano. Furthermore, thanks to the strong effort spent on defining the fields of the database, the structure proposed in this paper is open to the collection of further data coming from future improvements in the knowledge of the fault systems. By layering additional user-specific geographic information and managing the proposed database (topological querying) a great diversity of hazard and vulnerability maps can be produced by the user. This is a proposal of a backbone for a comprehensive geographical database of fault systems, universally applicable to other sites.

An overview of experimental models to understand a complex volcanic instability: Application to Mount Etna, Italy

Mon, 31 Dec 2012 23:00:00 GMT

Title: An overview of experimental models to understand a complex volcanic instability: Application to Mount Etna, Italy Authors: Acocella, V.; Dipartimento di Scienze Geologiche, Università Roma tre, Roma, Italy; Neri, M.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Catania, Catania, Italia; Norini, G.; Istituto per la Dinamica dei Processi Ambientali, Consiglio Nazionale delle Ricerche, Dalmine, Italy Abstract: Volcanic edifices are often unable to support their own load, triggering the instability of their flanks. Many analogue models have been aimed, especially in the last decade, at understanding the processes leading to volcano flank instability; general behaviors were defined and the experimental results were compared to nature. However, available data at well-studied unstable volcanoes may allow a deeper understanding of the specific processes leading to instability, providing insights also at the local scale. Etna (Italy) constitutes a suitable example for such a possibility, because of its well-monitored flank instability, for which different triggering factors have been proposed in the last two decades. Among these factors, recent InSAR data highlight the role played by magmatic intrusions and a weak basement, under a differential unbuttressing at the volcano base. This study considers original and recently published experimental data to test these factors possibly responsible for flank instability, with the final aim to better understand and summarize the conditions leading to flank instability at Etna. In particular, we simulate the following processes: a) the longterm activity of a lithospheric boundary, as the Malta Escarpment, separating the Ionian oceanic lithosphere from the continental Sicilian lithosphere, below the most unstable east flank of the volcano; b) spreading due to a weak basement, with different boundary conditions; c) the pressurization of a magmatic reservoir, as that active during the 1994–2001 inflation period; d) dike emplacement, as observed during the major 2001 and 2002–2003 eruptions. The experimental results suggest that: 1) the long-term activity of a lithospheric tectonic boundary may create a topographic slope which provides a differential buttressing at the volcano base, a preparing factor to drive longer-term (>105 years) instability on the east flank of the volcano; 2) volcano spreading (b104 years) has limited effect on flank instability at Etna; 3) magmatic intrusions (b101 years), both in the form of Mogi-like sources or dikes, provide the most important conditions to trigger flank instability on the shorter-term.

Analytical representation of the fault slip velocity from spontaneous dynamic earthquake models

Thu, 31 May 2012 22:00:00 GMT

Title: Analytical representation of the fault slip velocity from spontaneous dynamic earthquake models Authors: Bizzarri, A.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Bologna, Bologna, Italia Abstract: We have analyzed the most relevant features of three different analytical representations of the time evolution of the cosesimic slip velocity derived from theoretical basis; the so-called modified Yoffe function (MY), which pertains to a singular crack solution, the solution for a nonspontaneous crack obeying a position-weakening governing equation (PR) and the solution for a 1-D fault model subject to a linear slip-weakening friction law (B). By considering the same input parameters, we quantitatively compare these slip velocity functions (SVF) and we found that the time evolutions of the velocity and the correspondent slip predicted by the MY and B functions are very similar, while the PR predicts a very sharp peak. Correspondingly, the PR SVF is richer in high frequency and the fall off of its spectrum at high frequencies goes roughly as w–1.5, while those of MY and B more closely follow w–2. Then we select two spontaneous, 3-D, dynamic, subshear models, representing a crack-like or a pulse-like rupture and we account for both homogeneous and heterogeneous configurations. We then compare the three SVF in order to see how they are able to reproduce the 3-D solutions; we also show how the input parameters of the SVF can be constrained from the results of the dynamic models. In the homogeneous cases our results indicate that the MY and the PR SVF reproduce adequately well the main features of a dynamic solution in the case of a crack-like rupture. The PR function overestimates vpeak and the MY SVF predicts a too rapid deceleration. In the case of a pulse-like rupture both the MY and the B SVF tend to underestimate vpeak , but all of them capture very well the final cumulated fault slip. Moreover, the B function fits better that the MY the overall behavior of the fault slip. The considered SVF are able to reproduce the spectral fall off of a 3-D solution at intermediate frequencies (for w < 20 Hz), the MY and the PR for a crack-like rupture and the MY and the B SVF for a pulse-like rupture. In particular, for w < 10 Hz the spectral content of the B function is practically indistinguishable from that of the spontaneous pulse-like solution. In the heterogeneous configurations the analytical functions cannot reproduce all the spectral details of the numerical solutions, but we see how it is possible to fit the overall behavior of a single pulse in fault the slip velocity time history. The thorough analysis performed in this work can contribute to the discussion about the debated choice of the source time function to be used in the kinematic models, which in turn is extremely important in the contest of hazard assessment and ground motions generation, although stress heterogeneities, geometrical irregularities, attenuation and free surface effects can definitively smear the details of the analytical functions.

crustal fracturing field and presence of fluid as revealed by seismic anisotropy: case-histories from seismogenic areas in the Apennines

Wed, 16 Feb 2011 23:00:00 GMT

Title: crustal fracturing field and presence of fluid as revealed by seismic anisotropy: case-histories from seismogenic areas in the Apennines Authors: Pastori, Marina; Università degli studi di Perugia Abstract: During the last decades, the study of seismic anisotropy has provided useful information for the interpretation and evaluation of the stress field and active crustal deformation. Seismic anisotropy can yield valuable information on upper crustal structure, fracture field, and presence of fluid-saturated rocks crossed by shear waves. Several studies worldwide demonstrate that seismic anisotropy is related to stress-aligned, filled-fluid micro-cracks (EDA model, Crampin et al., 1984b; Crampin, 1993). The seismic anisotropy is an almost ubiquitous property of the Earth and the Shear Wave Splitting is the most unambiguous indicator of anisotropy, but the automatic estimation of the splitting parameters is difficult because the effect of the anisotropy on a seismogram is a second order, not easily detectable effect. Different researchers developed automated techniques aimed to study the Shear Wave Splitting: in this study, the results of different codes are compared in order to evaluate the best method for automatic anisotropy evaluation. In the last three years, an automatic analysis code, “Anisomat+”, was developed, tested and improved to calculate the anisotropic parameters: fast polarization direction () and delay time (∂t). “Anisomat+” consists of a set of MatLab scripts able to retrieve automatically crustal anisotropy parameters from three-component seismic recordings of local earthquakes. It needs waveforms and hypocentral parameters in the format routinely archived by the Istituto Nazionale di Geofisica e Vulcanologia (INGV). The code uses horizontal component cross-correlation method: a mathematical algorithm aimed to measure the similarity of the pulse shape between two shear waves. Anisomat+ has been compared to other two automatic analysis codes (SPY and SHEBA) and tested on three zones of the Apennines (Val d’Agri, Tiber Valley and L’Aquila surroundings). It was observed that, if the number of measures is large enough, at each station the average values of the parameters (fast direction and delay time) are comparable. The main goal in developing of an automatic code was to have tool able to work on a big amount of data, in a short time, by reducing the errors due to the subjectivity. These two acquirements are very useful and are the basis to develop a quasi real-time monitoring of the anisotropic parameters. The anisotropic parameters, resulting from the automatic computation, have been interpreted to determine the fracture field geometries; for each area, I defined the dominant fast direction and the intensity of the anisotropy, interpreting these results in the light of the geological and structural setting and of two anisotropic interpretative models, proposed in the literature. In the first one, proposed by Zinke and Zoback (2000), the local stress field and cracks are aligned by tectonics phases and are not necessarily related to the presently active stress field. Therefore the anisotropic parameters variations are only space-dependent. In the second, EDA model (Crampin, 1993), and its development in the APE model (Zatsepin and Crampin, 1995) fluid-filled micro-cracks are aligned or ‘opened’ by the active stress field and the variation of the stress field might be related to the evolution of the pore pressure in time; therefore in this case the variation of the anisotropic parameters are both space- and time- dependent. I recognized that the average of fast directions, in the three selected areas, are oriented NW-SE, in agreement with the orientation of the active stress field, as suggested by the EDA model, proposed by Crampin (1993), but also, by the proposed by Zinke and Zoback model; in fact, NW-SE direction corresponds also to the strike of the main fault structures in the three study regions. The mean values of the magnitude of the normalized delay time range from 0.005 s/km to 0.007 s/km and to 0.009 s/km, respectively for the L'Aquila (AQU) area, the High Tiber Valley (ATF) and the Val d'Agri (VA), suggesting a 3-4% of crustal anisotropy (Piccinini et al., 2006). In each area are also examined the spatial and temporal distribution of anisotropic parameters, which lead to some innovative observations, listed below. o The higher values of normalized delay times have been observed in those zones where most of the seismic events occur. This aspect was further investigated, by evaluating the average seismic rate, in a time period, between years 2005 and 2010, longer than the lapse of time, analyzed in the anisotropic studies. This comparison has highlighted that the value of the normalised delay time is larger where the seismicity rate is higher. o In the Alto Tiberina Fault area the higher values of normalised delay time are not only related to the presence of a high seismicity rate but also to the presence of a tectonically doubled carbonate succession. Therefore, also the lithology, plays a important role in hosting and preserving the micro-fracture network responsible for the anisotropic field. o The observed temporal variations of anisotropic parameters, have been observed and related to the fluctuation of pore fluid pressure at depth possibly induced by different mechanisms in the different regions, for instance, changes in the water table level in Val D’Agri (Valoroso et al., GJI submitted), occurrence of the April 6th Mw=6.1 earthquake in L’Aquila (Lucente et al., 2010). Since these variations have been recognized, it is possible to affirm that the models that better fit my results, both in term of fast directions and of delay times, seems to be those proposed by Crampin (1993) and Zatsepin & Crampin (1995), respectively EDA and APE models.

Project S1: Analysis of the seismic potential in Italy for the evaluation of the seismic hazard

Tue, 29 Jun 2010 22:00:00 GMT

Title: Project S1: Analysis of the seismic potential in Italy for the evaluation of the seismic hazard Authors: Barba, S.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Roma, Italia; Doglioni, C.; Sapienza Università di Roma Abstract: The project S1 was aimed at (a) collecting new data and to update the existing databases needed to quantify seismic hazard; (b) promoting new studies on specific fields of knowledge and less-explored areas of Italy; (c) testing new approaches to evaluate seismic potential; (d) bounding slip rate values to use within probabilistic hazard estimates; and (e) preparing the way towards a future seismic hazard map of Italy. It was designed with three scientific parts – nationwide basic data, rheology, and field studies – and implemented into four tasks: 1) earthquake geodesy and modeling, 2) seismological data and earthquake statistics, 3) earthquake geology, and 4) tsunamis. Although with many difficulties and some delay, described in the appropriate section, all the above objectives have generally been accomplished. New observations were collected through original fieldwork and more sophisticated analyses were performed on existing data. Datasets needed for the seismic hazard estimates were updated at various levels by reducing both epistemic and aleatory uncertainties. New studies were carried out on specific fields of knowledge, e.g. addressing the repeatability of geodetic and stress data measurements or the seismogenic behavior of misoriented faults. Studies on less-explored areas were stimulated, and faults, whose seismic potential was not previously accounted for, were mapped and/or parameterized in the Ionian and Adriatic Seas, in Calabria, Sicily and the Southwestern Alps. Independent approaches to evaluate the seismic potential were tested, and a large effort toward homogenization and verifiability was made. The substantial improvements of nationwide datasets and understanding of the tectonic processes in large areas of the country set the basis for a significantly better assessment of seismic hazard.

Effects of permeability and porosity evolution on simulated earthquakes

Mon, 30 Apr 2012 22:00:00 GMT

Title: Effects of permeability and porosity evolution on simulated earthquakes Authors: Bizzarri, A.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Bologna, Bologna, Italia Abstract: Numerical simulations are a fundamental tool to access the typical conditions attained during earthquake instabilities and to simulate the large number of dissipative processes taking places during faulting. In this study we consider a single-degree-of-freedom spring-slider system, a simplified fault model which can describe the whole seismic cycle and the dynamics of a fault with spatially homogeneous properties. We assume a rate- and state-dependent friction in which we incorporate the effects of pore fluid pressure, thermally-pressurized as a consequence of the frictional heat produced during sliding. We explore, in a single framework, the role of the time variations of the porosity, permeability or both, ultimately leading to changes in hydraulic diffusivity, which has been recognized as one of the key parameters in thermally-pressurized faults. Our synthetic ruptures show that the changes in the hydraulic diffusivity only due to porosity variations do not markedly affect the earthquake recurrence (cycle time), the traction evolution and the thermal history of the fault. On the contrary, when the evolutions of both the porosity and the permeability are accounted for, the cycle time is significantly reduced. This result has a clear implication in the context of the hazard assessment.

Principal Slip Zones in Limestone: Microstructural Characterization and Implications for the Seismic Cycle (Tre Monti Fault, Central Apennines, Italy)

Fri, 31 Dec 2010 23:00:00 GMT

Title: Principal Slip Zones in Limestone: Microstructural Characterization and Implications for the Seismic Cycle (Tre Monti Fault, Central Apennines, Italy) Authors: Smith, S.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Roma, Italia; Di Toro, G.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Roma, Italia; Speiss, R.; Dipartimento di Geoscienze, Universita` degli Studi di Padova, 1 Via Giotto, 35137 Padova, Italy; Billi, A.; Istituto di Geologia Ambientale e Geoingegneria, CNR, Via Salaria km 29.3, 00015 Monterotondo (Rome), Italy Abstract: Abstract—Earthquakes in central Italy, and in other areas worldwide, often nucleate within and rupture through carbonates in the upper crust. During individual earthquake ruptures, most fault displacement is thought to be accommodated by thin principal slip zones. This study presents detailed microstructural observations of the slip zones of the seismically active Tre Monti normal fault zone. All of the slip zones cut limestone, and geological constraints indicate exhumation from\2 km depth, where ambient temperatures are 100 C. Scanning electron microscope observations suggest that the slip zones are composed of 100% calcite. The slip zones of secondary faults in the damage zone contain protocataclastic and cataclastic fabrics that are cross-cut by systematic fracture networks and stylolite dissolution surfaces. The slip zone of the principal fault has much more microstructural complexity, and contains a 2–10 mm thick ultracataclasite that lies immediately beneath the principal slip surface. The ultracataclasite itself is internally zoned; 200–300 lm-thick ultracataclastic sub-layers record extreme localization of slip. Syn-tectonic calcite vein networks spatially associated with the sub-layers suggest fluid involvement in faulting. The ultracataclastic sub-layers preserve compelling microstructural evidence of fluidization, and also contain peculiar rounded grains consisting of a central (often angular) clast wrapped by a laminated outer cortex of ultra-fine-grained calcite. These ‘‘clast-cortex grains’’ closely resemble those produced during layer fluidization in other settings, including the basal detachments of catastrophic landslides and saturated high-velocity friction experiments on clay-bearing gouges. An overprinting foliation is present in the slip zone of the principal fault, and electron backscatter diffraction analyses indicate the presence of a weak calcite crystallographic preferred orientation (CPO) in the fine-grained matrix. The calcite c-axes are systematically inclined in the direction of shear. We suggest that fluidization of ultracataclastic sub-layers and formation of clast-cortex grains within the principal slip zone occurred at high strain rates during propagation of seismic ruptures whereas development of an overprinting CPO occurred by intergranular pressure solution during post-seismic creep. Further work is required to document the range of microstructures in localized slip zones that cross-cut different lithologies, and to compare natural slip zone microstructures with those produced in controlled deformation experiments.

Low- to high-velocity frictional properties of the clay-rich gouges from the slipping zone of the 1963 Vaiont slide, northern Italy

Fri, 31 Dec 2010 23:00:00 GMT

Title: Low- to high-velocity frictional properties of the clay-rich gouges from the slipping zone of the 1963 Vaiont slide, northern Italy Authors: Ferri, F.; Univ. di Padove, Dpt. Geoscienze; Di Toro, G.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Roma, Italia; Han, R.; Korea Institute of Geoscience and Mineral Resources, Daejeon, South Korea; Noda, H.; Seismological Laboratory, California Institute of Technology, Pasadena, California, USA; Shimamoto, T.; State Key Laboratory of Earthquake Dynamics, Institute of Geology, Chinese Earthquake Administration, Beijing, China; Quaresimin, M.; Dipartimento di Ingegneria dei Sistemi Industriali, Università degli Studi di Padova, Padua, Italy; De Rossi, N.; Dipartimento di Ingegneria dei Sistemi Industriali, Università degli Studi di Padova, Padua, Italy; Hirose, T.; Kochi Institute for Core Sample Research, JAMSTEC, Kochi, Japan Abstract: The final slip of about 450 m at about 30 m/s of the 1963 Vaiont landslide (Italy) was preceded by >3 year long creeping phase which was localized in centimeter-thick clay-rich layers (60–70% smectites, 20–30% calcite and quartz). Here we investigate the frictional properties of the clay-rich layers under similar deformation conditions as during the landslide: 1–5 MPa normal stress, 2 × 10−7 to 1.31 m/s slip rate and displacements up to 34 m. Experiments were performed at room humidity and wet conditions with biaxial, torsion and rotary shear apparatus. The clay-rich gouge was velocity-independent to velocity-weakening in both room humidity and wet conditions. In room humidity experiments, the coefficient of friction decreased from 0.47 at v < 5.0 × 10−5 m/s to 0.12 at 1.31 m/s. Microstructural and mineralogical analyses of the gouge after experiments indicate that the dramatic weakening results from thermo-chemical pressurization of pore fluids (smectite decomposition to illite-type clays) and powder lubrication. In wet experiments, the coefficient of friction decreased from 0.17 at v < 1.0 × 10−4 m/s to 0.0 at v > 0.70 m/s: full lubrication results from the formation of a continuous water film in the gouge. The Vaiont landslide occurred under wet to saturated conditions. The unstable behavior of the landslide is explained by the velocity-weakening behavior of the Vaiont clay-rich gouges. The formation of a continuous film of liquid water in the slipping zone reduced the coefficient of friction to almost zero, even without invoking the activation of thermal pressurization. This explains the extraordinary high velocity achieved by the slide during the final collapse.

Fault roughness at seismogenic depths from LIDAR and photogrammetric analysis

Fri, 31 Dec 2010 23:00:00 GMT

Title: Fault roughness at seismogenic depths from LIDAR and photogrammetric analysis Authors: Bistacchi, A.; Univ. Milano Bicocca, Dpt. Geologia; Griffith, W. A.; Smith, S.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Roma, Italia; Di Toro, G.; Jones, R.; Geospatial Research Ltd., Department of Earth Sciences, University of Durham, Durham DH1 3LE, UK; Nielsen, S.; Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma1, Roma, Italia Abstract: Abstract—Fault surface roughness is a principal factor influencing earthquake mechanics, and particularly rupture initiation, propagation, and arrest. However, little data currently exist on fault surfaces at seismogenic depths. Here, we investigate the roughness of slip surfaces from the seismogenic strike-slip Gole Larghe Fault Zone, exhumed from ca. 10 km depth. The fault zone exploited pre-existing joints and is hosted in granitoid rocks of the Adamello batholith (Italian Alps). Individual seismogenic slip surfaces generally show a first phase of cataclasite production, and a second phase with beautifully preserved pseudotachylytes of variable thickness. We determined the geometry of fault traces over almost five orders of magnitude using terrestrial laser-scanning (LIDAR, ca. 500 to\1 m scale), and 3D mosaics of high-resolution rectified digital photographs (10 m to ca. 1 mm scale). LIDAR scans and photomosaics were georeferenced in 3D using a Differential Global Positioning System, allowing detailed multiscale reconstruction of fault traces in Gocad . The combination of LIDAR and high-resolution photos has the advantage, compared with classical LIDARonly surveys, that the spatial resolution of rectified photographs can be very high (up to 0.2 mm/pixel in this study), allowing for detailed outcrop characterization. Fourier power spectrum analysis of the fault traces revealed a self-affine behaviour over 3–5 orders of magnitude, with Hurst exponents ranging between 0.6 and 0.8. Parameters from Fourier analysis have been used to reconstruct synthetic 3D fault surfaces with an equivalent roughness by means of 2D Fourier synthesis. Roughness of pre-existing joints is in a typical range for this kind of structure. Roughness of faults at small scale (1 m to 1 mm) shows a clear genetic relationship with the roughness of precursor joints, and some anisotropy in the selfaffine Hurst exponent. Roughness of faults at scales larger than net slip ([1–10 m) is not anisotropic and less evolved than at smaller scales. These observations are consistent with an evolution of roughness, due to fault surface processes, that takes place only at scales smaller or comparable to the observed net slip. Differences in roughness evolution between shallow and deeper faults, the latter showing evidences of seismic activity, are interpreted as the result of different weakening versus induration processes, which also result in localization versus delocalization of deformation in the fault zone. From a methodological point of view, the technique used here is advantageous over direct measurements of exposed fault surfaces in that it preserves, in cross-section, all of the structures which contribute to fault roughness, and removes any subjectivity introduced by the need to distinguish roughness of original slip surfaces from roughness induced by secondary weathering processes. Moreover, offsets can be measured by means of suitable markers and fault rocks are preserved, hence their thickness, composition and structural features can be characterised, providing an integrated dataset which sheds new light on mechanisms of roughness evolution with slip and concomitant fault rock production.