# Engineering geological zonation of a complex landslide system through seismic ambient noise measurements at the Selmun Promontory (Malta)

Engineering geological zonation of a complex landslide system through seismic ambient noise... Summary The cliff slope of the Selmun Promontory, located in the Northern part of the island of Malta (Central Mediterranean Sea) close to the coastline, is involved in a landslide process as exhibited by the large block-size talus at its bottom. The landslide process is related to the geological succession outcropping in the Selmun area, characterized by the overposition of a grained limestone on a plastic clay, that induces a lateral spreading phenomenon associated with detachment and collapse of different-size rock blocks. The landslide process shapes a typical landscape with a stable plateau of stiff limestone bordered by an unstable cliff slope. The ruins of Għajn Ħadid Tower, the first of the 13 watchtowers built in 1658 by the Grand Master Martin de Redin, stand out on the Selmun Promontory. The conservation of this important heritage site, already damaged by an earthquake which struck the Maltese Archipelago on 1856 October 12, is currently threatened by a progressive retreat of the landslide process towards the inland plateau area. During 2015 and 2016, field surveys were carried out to derive an engineering geological model of the Selmun Promontory. After a high-resolution geomechanical survey, the spatial distribution of the joints affecting the limestone was obtained. At the same time, 116 single-station noise measurements were carried out to cover inland and edge of the limestone plateau as well as the slope where the clays outcrop. The obtained 1-hour time histories were analysed through the horizontal to vertical spectral ratio technique, as well as polarization and ellipticity analysis of particle motion to define the local seismic response in zones having different stability conditions, that is, related to the presence of unstable rock blocks characterized by different vibrational modes. The results obtained demonstrate the suitability of passive seismic geophysical techniques for zoning landslide hazard in case of rock slopes and prove the relevance of anisotropies in conditioning the polarization of vibrational modes for dislodged rock masses. Geomechanics, Geomorphology, Seismic noise, Site effects 1 INTRODUCTION Landslides are one of the major natural hazards, causing large human and economic losses worldwide (Petley 2012). In coastal areas, landslides and related risk have been widely studied in the scientific community due to the presence of populated settlements as well as touristic and cultural heritage sites (Guzzetti et al. 1999; Senfaute et al. 2009; Pennetta & Russo 2011; Montoya-Montes et al. 2012; Dewez et al. 2013; Epifanio et al. 2013; Martino & Mazzanti 2014). Commonly, sea cliffs are subjected to retreat during their complex evolution, governed by a combination of sea erosion and gravity-induced instability processes on coastal cliff slopes (Hampton et al. 2004; Young & Ashford 2008; Dickson et al. 2013; Rosser et al. 2013). Such processes may produce landslides extending from a single sea cliff up to several hundred meters of coast (Hutchinson et al. 1991; Della Seta et al. 2013). The island of Malta (Central Mediterranean Sea) presents several significant case studies regarding coastal gravity- and erosion-induced instability processes. The geological setting of the Western region of the Maltese Archipelago, that is, the over-position of stiff limestone (Upper Coralline Limestone, UCL) on plastic clay deposits (Blue Clay, BC), leads to large lateral spreading processes associated with falls, slides and/or topples of different-size rock blocks. These instabilities affect both inland areas, for example, the plateaus on which the cities of Mdina (Malta) and the Citadel of Victoria (Gozo) are built (Gigli et al. 2012), as well as sea cliffs all along the coastline, especially in the northwestern part of Malta and in Gozo (Mantovani et al. 2013; Galea et al. 2014). In recent years, multidisciplinary approaches have been developed to investigate landslides by coupling traditional engineering geological surveying (ISRM 1978) with geophysical techniques (Bogoslovsky & Ogilvy 1977; McCann & Forster 1990; Hack 2000; Jongmans & Garambois 2007; Maurer et al. 2010) that present advantages such as flexibility, efficiency, ease of instrument deployment and rapidity of data processing. In general, geophysical methods are applied to provide a 2-D or 3-D imaging of landslide mass (e.g. dimensions and geometry) and some basic features (e.g. water content or seismic wave velocity), as evidenced by published works in which different approaches were applied to investigate landslide, for example, electrical resistivity tomography (Havenith et al. 2000; Hibert et al. 2012), spontaneous potential (Chambers et al. 2011), ground penetrating radar (Bichler et al. 2004; Sass et al. 2008), gravimetry (Del Gaudio et al. 2000), electromagnetism (Méric et al. 2005), seismic reflection (Bichler et al. 2004), seismic refraction (Havenith et al. 2000; Glade et al. 2005), seismic tomography (Meric et al. 2005) and Multi-channel Analysis of Surface Waves (Imposa et al. 2017). Passive seismic surveys and, in particular, seismic ambient noise measurements have already been applied to investigate landslide-affected slopes (Del Gaudio et al. 2008; Burjánek et al. 2010, 2012; Yalcinkaya et al. 2016; Del Gaudio 2017; Pazzi et al. 2017), unstable blocks of rocky cliffs (Got et al. 2010; Lévy et al. 2010, 2011; Panzera et al. 2012; Bottelin et al. 2013a,b, 2014, 2017; Galea et al. 2014; Colombero et al. 2017) and natural rock arches (Starr et al. 2015; Moore et al. 2016) by different approaches, for example, horizontal to vertical spectral ratio (HVSR) analysis, f–k analysis, site to reference spectral ratios, polarization analysis, noise spectral and amplitude variations. With respect to other geophysical methods, the published applications of passive seismic approaches demonstrated their capability to take into account the time dimension, with the aim of observing the variation of specific effects or parameters with time. Such a capability can be applied to characterize the dynamic behaviour of a landslide in terms of resonance frequency of the unstable mass (Got et al. 2010; Lévy et al. 2010; Bottelin et al. 2013a,b, 2017; Valentin et al. 2017). According to the published results, a joint application of HVSR analysis and polarization analysis appears to yield reliable estimates of the dimension of the landslide body and to characterize its seismic response (Burjánek et al. 2010, 2012; Galea et al. 2014). Characterization of the vibrational behaviour of landslide affected slopes can be considered a useful tool to design strategies for managing landslide risk, for example, for planning and monitoring consolidation intervention (Bottelin et al. 2017) or designing monitoring systems as already experienced in the case of drainage plants (Lenti et al. 2012; Fiorucci et al. 2017), hydroelectric power stations (Tang et al. 2015; Xu et al. 2015) and geothermal heat exchangers (Zang et al. 2017). In the light of the perspectives demonstrated by seismic noise measurements, in this work we investigate the rock cliff slope of the Selmun Promontory, located on the northwestern coast of Malta and affected by an important landslide process. We apply a joint approach, including HVSR analysis and polarization analysis. Such an approach was applied to study the local seismic response and the vibrational behaviour of the cliff slope as well as to evaluate its usefulness as a tool in the framework of landslide hazard zonation, that is, the procedure for dividing the land surface into homogeneous areas, and ranking of these areas according to their degree of stability (Varnes & IAEG 1984). The Selmun Promontory was chosen as a case study also because the ruins of Għajn Ħadid Tower, the first of the thirteen watchtowers built in 1658 by the Grand Master Martin de Redin, stand out prominently in this area. Currently the conservation of this important heritage site, already damaged by an earthquake that struck the Maltese Archipelago on 1856 October 12, is threatened by the continuous deformations related to the spreading process which is evolving over the years from the coastline towards the inland plateau area. In 2015 and 2016, a detailed engineering geological survey was carried out to define the geological setting of the Selmun Promontory and to obtain the mechanical properties of the jointed rock mass. At the same time, 116 single-station seismic noise measurements were carried out over a wide area on both unstable and stable regions. This allowed the characterization of the local seismic response in terms of zones having different stability conditions, mainly related to the presence of already isolated rock blocks with respect to adjacent stable zones. The obtained results represent a fundamental contribution in the framework of designing protection strategies for managing the landslide process in the Selmun Promontory, that is, for ensuring the safety of tourists and preserving the heritage site of Għajn Ħadid Tower. 2 GEOLOGICAL AND GEOMORPHOLOGICAL SETTINGS The Maltese Archipelago is composed of three main islands (Malta, Gozo and Comino) located in the middle of the Mediterranean Sea, about 100 km South from Sicily and 290 km East from Tunisia. The islands represent the only currently emergent part of an extensive shallow-water shelf that extends from Eastern Sicily to the Malta Graben, an important part of the threshold separating the Western and Eastern Mediterranean basins (Pedley 2011). Limestones and clays of Oligocene and Miocene epochs form the Maltese islands. The sequence of the marine sedimentary rocks of the Maltese Archipelago (Fig. 1) is composed of five main geological formations (Hyde 1955; Pedley et al. 1976, 1978, 2002; Gigli et al. 2012) whose oldest formation is Lower Coralline Limestone (LCL) of Oligocene age (Chattian). Figure 1. View largeDownload slide Geology of the Maltese Archipelago: (a) geology map, position of the Maltese islands in the Mediterranean Sea and location of the Selmun Promontory in Malta (in the red frame); (b) sketch of the sedimentary sequence. Figure 1. View largeDownload slide Geology of the Maltese Archipelago: (a) geology map, position of the Maltese islands in the Mediterranean Sea and location of the Selmun Promontory in Malta (in the red frame); (b) sketch of the sedimentary sequence. Above this limestone, three formations are present and outcrop at the Selmun Promontory (Fig. 2a). Starting from the lowermost layer, these are: Globigerina Limestone (GL) formation, a soft yellowish fine-grained limestone of Lower Miocene age (Aquitanian-Langhian); BC formation, a soft pelagic blue or greenish grey marl and limey clay of Middle Miocene age (Serravallian); UCL formation, pale grey and orange fossiliferous coarse-grained limestone, of Upper Miocene in age (Tortonian-Messinian). Figure 2. View largeDownload slide Geological and geomorphological setting of the Selmun Promontory: (a) Photograph showing the UCL cliff slope, the UCL–BC and BC–GL stratigraphic geological contacts and the debris covering the BC slope; (b–d) Photographs showing the evolution over time of the landslide process, from the development of fractures in the UCL plateau (left) to the detachment of UCL rock blocks (right). Figure 2. View largeDownload slide Geological and geomorphological setting of the Selmun Promontory: (a) Photograph showing the UCL cliff slope, the UCL–BC and BC–GL stratigraphic geological contacts and the debris covering the BC slope; (b–d) Photographs showing the evolution over time of the landslide process, from the development of fractures in the UCL plateau (left) to the detachment of UCL rock blocks (right). Based on the geological setting, the island of Malta can be divided in two main sectors (Fig. 1a): the southeastern part of Malta where only the two oldest formations outcrop, shaping relatively flat landscapes, and the northwestern part of the island, where the full sedimentary sequence is generally conserved and the UCL and the BC mainly outcrop. Due to the outcropping geological formations, unstable cliff slopes typically forming the edges of a summit UCL plateau overtopping gentle slopes of the BC (between 30° and 45°) form the typical landscapes of the northwestern part of Malta (Gigli et al. 2012; Mantovani et al. 2013; Galea et al. 2014). The Selmun Promontory, located on the North-facing coast of the northwestern sector of Malta, was chosen as a case study because it is involved in a significant landslide process, as shown by the large block-size talus and the dense joint net clearly visible at the top of the stiff plateau (Fig. 2). In particular, the geological succession of the Selmun Promontory, characterized by the overposition of stiff rocks on a plastic deposit (i.e. the stiff UCL on the plastic BC), leads to a lateral spreading phenomenon (Goudie 2004): the horizontal deformation affecting the clayey materials, having viscoplastic behaviour, induces fracturing of the overlying stiff rock. Lateral spreading shapes a plateau of stiff rock bordered by jointed unstable cliffs, favouring the detachment of single rock blocks by typical gravity-induced instability mechanisms, that is, planar sliding, wedge sliding, toppling and falling (Hoek & Bray 1981). The resulting landslide process should be defined as complex type according to Varnes (1978) and Hutchinson (1988). During 2015 and 2016, field activities and satellite image studies were carried out to better characterize the geological features of the Selmun Promontory. The stratigraphic succession of the Selmun Promontory (Fig. 2a) is composed of 20–30 m of the UCL, 30 m of the BC and the GL at the base. All the formations show almost horizontal or slightly NE-dipping (<5°) strata. On the gentle slope, both the BC and the GL rarely outcrop because a significant slope debris deposit resulting from the UCL plateau evolution generally covers them. The observed slope debris is composed of UCL clasts having heterogeneous sizes (from centimetre-scale clasts up to metre-size blocks) embedded in weathered BC and residual material of limestone dissolution. The slope debris deposits reach several meters of thickness, especially in the part close to the UCL cliff wall where the largest-size rock blocks are located. By considering the Selmun Promontory morphology, it can be schematized as suggested by Martino & Mazzanti (2014) for coastal slopes as the final stage of evolution of a sea cliff, that is, a cliff no longer directly affected by sea erosion (except during exceptional storm events) that retreats primarily by gravity-induced instability processes; at this stage, a beach separates the old sea cliff from the current shoreline. 3 ENGINEERING GEOLOGICAL SURVEY Detailed engineering geological field surveys were carried out to define the geomechanical properties of the jointed rock mass composing the cliff slope of the Selmun Promontory (Iannucci et al. 2017). First of all, the joints recognized in the rock mass (on the plateau surface as well as on the cliff wall) were mapped and plotted on a satellite view combining information derived by a GPS device and field observations, such as direction and length of joint segments (with a decimetre resolution). Then, according to the ISRM standard (ISRM 1978), each joint was characterized by a number of features, such as attitude (dip direction and dip) that allowed to define two main systems of sub-vertical joints at the Selmun Promontory: J1 prevalent in the NW zone (mean dip direction 330°, strike 60°) and J2 prevalent in the SE one (mean dip direction 45°, strike 135°). The two joint systems have a dip direction approximately orthogonal to the plateau edge direction and, therefore, a strike similar to it (Fig. 3). A GIS geo-database was implemented to inventory all the collected field data. Additionally, 18 cubic samples of the UCL were analysed through laboratory tests to characterize the rock mass matrix and allowed us to obtain, through the hydrostatic weighing method (ISRM 1979), a density (ρ) value of 2146 kg m−3 with a standard deviation of 141 kg m−3. Table 1 summarizes the main geomechanical parameters obtained for the surveyed joints in the rock mass of the Selmun Promontory. Figure 3. View largeDownload slide Results of the geomechanical survey performed at the Selmun Promontory: (a) Stereographic projection (equal-angle lower-hemisphere) of joint poles surveyed and planes of the two main joint systems; (b) Satellite view showing the location of the two main joint systems. Figure 3. View largeDownload slide Results of the geomechanical survey performed at the Selmun Promontory: (a) Stereographic projection (equal-angle lower-hemisphere) of joint poles surveyed and planes of the two main joint systems; (b) Satellite view showing the location of the two main joint systems. Table 1. Summary of the main geomechanical properties of the rock mass joints derived from field activities at the Selmun Promontory. Rock mass joints  ID system  Dip direction and dip (°)  Opening (cm)  Vertical offset (cm)  Hydraulic conditions  JCS (MPa)  JRC  J1  330/89  0–105 average 22  0–40 average 12  No flow  41.0  8–10  J2  45/88  0–70 average 16  0–30 average 16  No flow  41.0  8–10  Rock mass joints  ID system  Dip direction and dip (°)  Opening (cm)  Vertical offset (cm)  Hydraulic conditions  JCS (MPa)  JRC  J1  330/89  0–105 average 22  0–40 average 12  No flow  41.0  8–10  J2  45/88  0–70 average 16  0–30 average 16  No flow  41.0  8–10  View Large Table 2. Summary of the ambient noise measurements carried out in the several zones over the Selmun area (see Fig. 4). ID zone  Area (m2)  Measurements (n)  Measurement density (n m−2)  A  27 000  35  0.0013  B  1000  50  0.05  C  600  10  0.017  D  450  6  0.014  E  100 000  15  0.00015  ID zone  Area (m2)  Measurements (n)  Measurement density (n m−2)  A  27 000  35  0.0013  B  1000  50  0.05  C  600  10  0.017  D  450  6  0.014  E  100 000  15  0.00015  View Large An engineering geological model of the Selmun Promontory was obtained by combining the collected geological and geomechanical data and by mapping the spatial distribution of the joint net (Fig. 4); from this model, two orthogonal cross-sections of the Selmun Promontory were obtained (Fig. 5). Figure 4. View largeDownload slide Satellite view showing the geological setting of the Selmun Promontory: (1) Upper Coralline Limestone (Tortonian-Messinian); (2) Blue Clay (Serravallian); (3) Globigerina Limestone (Aquitanian-Langhian); (4) debris slope deposit (Quaternary); 5) attitude of strata; 6) fracture (the dashed line represents an inferred fracture); 7) Għajn Ħadid Tower; 8) seismic noise measurement station; 9) cross-section (see Fig. 5). Elevation is given by contour lines (m a.s.l.). The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 4. View largeDownload slide Satellite view showing the geological setting of the Selmun Promontory: (1) Upper Coralline Limestone (Tortonian-Messinian); (2) Blue Clay (Serravallian); (3) Globigerina Limestone (Aquitanian-Langhian); (4) debris slope deposit (Quaternary); 5) attitude of strata; 6) fracture (the dashed line represents an inferred fracture); 7) Għajn Ħadid Tower; 8) seismic noise measurement station; 9) cross-section (see Fig. 5). Elevation is given by contour lines (m a.s.l.). The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 5. View largeDownload slide Cross-sections (see Fig. 4) of the Selmun Promontory: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) seismic noise measurement station along the cross-section. Figure 5. View largeDownload slide Cross-sections (see Fig. 4) of the Selmun Promontory: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) seismic noise measurement station along the cross-section. Based on the engineering geological features, five zones were distinguished in the area (Fig. 4): one zone located in the inland plateau (A), characterized by the absence of ground fractures and unstable blocks; three zones located at the edge of the plateau (B–D), where a high density of ground fractures and unstable blocks exists; one zone located on the gentle slope (E), where the BC outcrops even if covered by debris. 4 SEISMIC NOISE ACQUISITION AND ANALYSIS During the 2015–2016 field activities, several geophysical campaigns of seismic noise measurements were carried out on the Selmun Promontory. Seismic ambient noise was recorded at 116 stations over an area of approximately 0.1 km2. Most of the noise measurement stations (Fig. 4) were distributed to cover the unstable zones of the cliff slope (zones B–D) as well as the stable UCL plateau (zone A), while a few measurements were carried out on the BC gentle slope (zone E). Details of the measurement campaigns are shown in Table 2. A three-component seismometer was deployed at each station: 84 measurements were carried out using an LE-3D/5 s seismometer by Lennartz Electronic GmbH coupled with a REFTEK 130–01 datalogger, set to a sampling frequency of 250 Hz; 32 measurements were carried out using a Tromino three-component portable seismograph (www.tromino.eu), set to a 128 or 256 Hz sampling frequency. In both cases, the seismic noise was recorded for 1 hour. Seismic noise is composed of ambient vibrations of the ground produced by random and uncontrolled sources, natural or related to human activity, such as tides, sea waves striking the coasts, wind, trees, buildings, industrial machinery, road traffic, trains, human footsteps, etc. In general, authors who investigated the composition of seismic noise (Douze 1964, 1967; Toksöz & Lacoss 1968; Li et al. 1984; Horike 1985; Yamanaka et al. 1994; Bonnefoy-Claudet et al. 2006) assume that it consists mainly of surface waves. The HVSR analysis, proposed by Nogoshi & Igarashi (1970, 1971), was applied by Nakamura (1989) to seismic noise measurements to evaluate the fundamental frequency of a site (f0), especially where there is a marked impedance contrast due to the presence of low shear-wave velocity layers above the seismic bedrock. A resonance peak in the distribution of the HVSR values versus natural frequency can be interpreted both in terms of SH-wave resonance in soft surface layers as well as in terms of the ellipticity of particle motion when the ambient noise wave train is made up predominantly of surface waves (Bonnefoy-Claudet et al. 2006). The wavefield is expected to be a combination of both types and the HVSR curve contains information about the shear wave velocity profile in shallow sediments (Galea et al. 2014). In this study, the HVSR analysis was carried out by using the open source software Geopsy (www.geopsy.org). Each recorded time history was divided into nonoverlapping windows of 40 s and the fast Fourier transform (FFT) was computed for the three spatial components ( north–south, east–west and up–down) in the frequency range between 1.0 and 60.0 Hz, using a 5 per cent cosine taper to each window. The Fourier spectra of each window were smoothed by the Konno–Ohmachi function (Konno & Ohmachi 1998). By taking the logarithmic average of the two horizontal components and averaging over the windows, the amplitude spectra and the H/V spectral ratio were finally computed for each single record. Vidale (1986) introduced in the seismological community the concept of polarization of the particle motion and a method to analyse it based on principal component analysis of the coherency matrix (i.e. complex covariance matrix), which is computed from analytical signals of a three-component seismogram and without time averaging. Based on the hypothesis that the wavefield within an unstable rock mass is dominated by normal mode vibration rather than horizontal propagation of seismic waves, Burjánek et al. (2010, 2012) implemented such an analysis by adopting a Continuous Wavelet Transform (CWT) in order to carry out a time–frequency domain analysis on a time history. WAVEPOL package (Burjánek et al. 2012) allows to obtain a 3-D ellipse representing the particle motion at each time-frequency pair, in a visual output expressing the frequency dependence of ellipticity and polarization. The ellipticity of the particle motion is defined as the ratio between the semi-minor axis and the semi-major axis of the ellipse (i.e. 1 represents circular particle motion and 0 purely linear motion) for each frequency. Two angle values indicate the polarization of the particle motion at each frequency: a strike value, representing the azimuth of the semi-major axis projected to the horizontal plane from North, and a dip value, describing the dip angle of the semi-major axis from the horizontal plane. Such a polarization analysis provides quite robust results about directivity effects of the particle motion, overcoming the bias that could be introduced by the denominator spectrum in the HVSR calculation (Panzera et al. 2014). In particular, by combining the outputs from such an analysis, that is, observing for the same frequencies a high degree of linearity on the ellipticity diagram and an azimuthal direction on the polar strike plot, it is possible to demonstrate polarization effects of the particle motion. 5 RESULTS AND DISCUSSION 5.1 HVSR analysis The results obtained from the HVSR analysis were grouped according to the five zones previously defined in the Selmun area (Fig. 6), in order to distinguish the vibrational behaviour related to engineering geological features. Figure 6. View largeDownload slide HVSR(f) functions grouped for different zones of the Selmun Promontory (see Fig. 4): the red curves represent examples of obtained HVSR(f) among the 116 measures, the solid black line is the average HVSR(f) function obtained by all curves in each zone and the dashed black lines show the standard deviation of all curves in each zone. Figure 6. View largeDownload slide HVSR(f) functions grouped for different zones of the Selmun Promontory (see Fig. 4): the red curves represent examples of obtained HVSR(f) among the 116 measures, the solid black line is the average HVSR(f) function obtained by all curves in each zone and the dashed black lines show the standard deviation of all curves in each zone. All the stations located on the UCL plateau (i.e. both inland plateau and on the fractured plateau edge) show HVSR curves characterized by a ubiquitous resonance peak in a narrow frequency range of 1.5–2.0 Hz followed by a sharp dip of the spectral ratio. The HVSR peak at around 1.5–2.0 Hz can be considered representative of the fundamental frequency of the site (f0) since this peak is always present in the Maltese archipelago where the UCL–BC–GL sedimentary sequence is located, as evidenced by the studies carried out at Dingli, Mdina, Wardija, Xemxija, Mellieħa, Anchor Bay, Mgarr, Nadur, Rabat, Golden Bay and Bahrija (Panzera et al. 2012, 2013; Vella et al. 2013; Galea et al. 2014; Farrugia et al. 2016, 2017; Pischiutta et al. 2017). This HVSR feature (peak at 1.5–2.0 Hz followed by a dip below 1.0 over a wide frequency range) was interpreted as related to the difference of seismic wave velocity values between the three geological formations that compose the sedimentary sequence. The dip of the HVSR(f) function after 2.0 Hz is presumably due to the shallow shear-wave velocity inversion at the interface between the competent UCL and the plastic BC, while the peak can be related to Rayleigh wave ellipticity and/or trapping of SH waves in the BC low-velocity layer (Galea et al. 2014). On the other hand, the HVSR(f) functions on the UCL plateau show marked differences in the seismic response between the unstable zones and the stable plateau area at frequencies higher than 3.0 Hz (Fig. 6). As a matter of fact, in the measurements carried out within and in proximity of the unstable zones, the HVSR curves show significant additional resonance peaks at higher frequency (3.0–60.0 Hz) that are not present in the measurements carried out on the stable plateau zone. In addition, peaks of the HVSR(f) functions are much more evident in areas having higher density of fractures and blocks, such as the jointed edge of the UCL plateau. Zone B shows a prevalent HVSR peak at frequency around 3.5 Hz, while the other jointed edge zones (C and D) do not show peculiar characteristic HVSR peaks. The HVSR(f) functions resulting from the noise measurements on the BC slope show a first main peak in a range between 2.4 and 6.5 Hz, therefore at higher frequencies than the stations located on the UCL plateau. This peak can be related to the resonance in the BC layer due to the lower values of the seismic wave velocity compared to the underlying GL layer. A rough evaluation of the fundamental frequency of a resonant layer (f0) can be theoretically obtained by the following relation (1):   $${f_0} = \frac{{{V_S}}}{{4h}},$$ (1)where VS is the S-wave velocity within the layer and h its thickness. The BC formation is characterized by an S-wave velocity of 300–400 m s−1 (Panzera et al. 2012; Farrugia et al. 2016, 2017) and its maximum thickness at the Selmun Promontory is 30 m. The seismic noise measurements were carried out on the debris deposit, about 1–5 m thick. For the measurements carried out in the higher part of the slope, we can assume a total thickness of the BC–debris layer of 35 m and a mean S-wave velocity of 350 m s−1, by considering also the poorer geomechanical features of the debris. Such parameters yield a fundamental frequency (f0) of about 2.5 Hz in the higher part of the slope, which increases to 5 Hz by halving the thickness of the BC–debris layer, a condition existing in the middle part of the slope. The obtained values of fundamental frequency (f0) are similar to those obtained by Panzera et al. (2012) at several sites on Malta where the BC outcrops, that is, Rabat, Mdina, Wardija and Mellieħa. The HVSR(f) functions in this region show additional peaks at higher frequency that can be related to the resonance in the debris deposit due to its local features (i.e. thickness and/or weathering), since most of the measurements on the BC clay were carried out on the debris that covers the slope (Fig. 4). Fig. 7 summarizes the HVSR results for the surveyed area, showing the spatial distribution of the frequency values and the HVSR(f) function amplitudes of the first HVSR peak. On the UCL plateau (zones A-D), the frequency value of the first peak of the HVSR(f) function is between 1.5 and 1.8 Hz. The variation of the main peak amplitude can be related to the location of the measurement points: the HVSR(f) function amplitude increases on moving from the inner plateau towards the edge of the plateau, where the joint density significantly increases. As an example, the Southern portion of zone A is characterized by HVSR(f) amplitude values between 2.0 and 4.0, while in zone B the HVSR(f) amplitude ranges from 6.0 to 10.0. Figure 7. View largeDownload slide Satellite view showing the measurement points (dots) and the values (in white) of the first peak of the HVSR(f) function (colour scale): (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) Għajn Ħadid Tower. The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 7. View largeDownload slide Satellite view showing the measurement points (dots) and the values (in white) of the first peak of the HVSR(f) function (colour scale): (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) Għajn Ħadid Tower. The light grey lines indicate the different zones of the Selmun Promontory (A–E). On the BC slope (zone E), the frequency of the first peak of the HVSR(f) function varies over a wider range compared to that on the UCL plateau, that is, between 2.4 and 6.5 Hz. Stations located in the higher part of the slope show lower resonance frequencies than the stations in the lower part. This variation can be explained by the decreasing thickness of BC on going down-slope, as shown in the engineering-geological cross-sections (Fig. 5). 5.2 Polarization analysis The polarization analysis confirms the peculiar features observed in the HVSR analysis and provides additional evidence to support the different seismic response in the various zones on the Selmun Promontory (Figs 8 and 9). Based on the polarization analysis, two features emerge with reference to specific frequency ranges: (i) the ellipse axes ratio, which decreases when the linearity of the particle motion increases; (ii) the azimuthal distribution of the particle motion, which reveals a main direction of polarization of the particle motion. Figure 8. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones A, B and C of the Selmun Promontory. Figure 8. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones A, B and C of the Selmun Promontory. Figure 9. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones D and E of the Selmun Promontory. Figure 9. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones D and E of the Selmun Promontory. By analysing all the results obtained on both the UCL plateau and the BC slope, the frequencies associated with the first peak of the HVSR(f) function (i.e. 1.5–1.8 Hz for the UCL plateau and 2.4–6.5 Hz for the BC slope) do not show evidence of polarization. As a matter of fact, even if the ellipticity diagrams show quite a high degree of linearity (an ellipse axes ratio in a 0.1–0.2 range), the polar strike plots do not demonstrate a main direction of the particle motion at these frequencies. This confirms the hypothesis that the first peak of HVSR(f) function is mainly related to the stratigraphic setting (i.e. 1-D resonance effect) and to the multi-directional nature of the ambient noise. On the inland UCL plateau (i.e. zone A), the noise analysis shows no evidence of either peaks of the HVSR(f) function (beyond the f0 peak) or polarization effects of the particle motion, except for a few cases where these effects can be observed at frequency values higher than 30 Hz and are probably related to very local elements (e.g. thin and shallow soil layers characterized by low Vs values). Zone B shows strong directivity effects at the same frequency of the 1-D resonance which results from the HVSR analysis. The polarization analysis shows that in zone B the 3.3–3.5 Hz frequency range is characterized by marked features of linearity (ellipse axes ratio 0.0–0.1) as well as by a strong azimuthal polarization of the particle motion. As a matter of fact, the polar strike plots show a directivity of the particle motion at azimuthal values between 135° and 150°, values similar to the dip direction of the joint system prevalent in zone B (J1). This result is analogous to those obtained by Burjánek et al. (2010) and Moore et al. (2011) at the landslide of Randa (Switzerland) as well as by Galea et al. (2014) at the unstable sea cliff of Anchor Bay (Malta). In zone D, the polarization effects occur along the azimuthal direction 45–50°, values similar to the dip direction of the main joint system of the area (J2). Nevertheless, these effects occur at two specific frequency values: 5 Hz closest to the cliff (0–5 m from the edge) and 7 Hz in the internal portion of the plateau edge (5–10 m from the edge). There is also evidence of polarization effects along a direction similar to the joint dip direction in zone C but not in a specific range of frequency. Finally, in zone E, the analysis does not show polarization effects, despite the several peaks observed in the HVSR(f) functions. Fig. 10 maps the results obtained through the polarization analysis and shows the spatial distribution of the first polarized frequency values (only for frequencies lower than 30 Hz) at the Selmun Promontory, the corresponding azimuth and the associated ellipticity values. Such a spatial distribution confirms the absence of polarized particle motion on the stable plateau (zone A), while highlighting different behaviours in the unstable areas (zones B–D). In particular, even though zone B is characterized by a denser joint net, its uniform seismic response leads us to consider a vibrational behaviour related to a unique rock block characterized by a single polarized frequency between 3.3 and 3.5 Hz. On the other hand, in zone D the seismic response seems to be related to two different, or independently vibrating rock blocks, as shown by the two polarized frequency values (5 and 7 Hz). Figure 10. View largeDownload slide Satellite view summarizing the results of polarization analysis for the first (lower value) polarized frequency: (1) UCL; (2) debris slope deposit; (3) fracture (the dashed line represents an inferred fracture); (4) Għajn Ħadid Tower. Each symbol indicates a noise measurement point: the label displays the value of the first polarized frequency; the arrow is oriented in the azimuthal direction of the polarization; the colour scale indicates the ellipticity intensity. Black dots indicate noise measurement points without polarization effects. The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 10. View largeDownload slide Satellite view summarizing the results of polarization analysis for the first (lower value) polarized frequency: (1) UCL; (2) debris slope deposit; (3) fracture (the dashed line represents an inferred fracture); (4) Għajn Ħadid Tower. Each symbol indicates a noise measurement point: the label displays the value of the first polarized frequency; the arrow is oriented in the azimuthal direction of the polarization; the colour scale indicates the ellipticity intensity. Black dots indicate noise measurement points without polarization effects. The light grey lines indicate the different zones of the Selmun Promontory (A–E). 6 ZONATION BASED ON VIBRATIONAL EFFECTS The different vibrational behaviour of the zones of the Selmun Promontory is shown in Fig. 11 where the results obtained for each noise station were projected and interpolated along the cross-section I–I’ (see Fig. 4 for location) with a gridding process using a triangulation with linear interpolation method. Additional seismic noise stations were projected on the cross-section to increase the density of results and improve the interpolation process: for the UCL plateau, we interpolated stations in a 20 m buffer from the track; for the BC slope we interpolated stations using an equal-altitude criteria, since the thicknesses variation of the BC layer and the debris deposit are negligible at the same altitude. Figure 11. View largeDownload slide Interpolation of the ambient noise measurement results along the cross-section I–I΄ (see Fig. 4) as a function of distance (x-axis) versus frequency (y-axis): (a) HVSR(f) amplitude; (b) ellipticity of the particle motion; (c) frequency of occurrence of polarized frequency; (d) cross-section I–I΄: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture; (6) seismic noise measurement station located just along the cross-section; (7) seismic noise measurement station projected on the cross-section (using a 20 m buffer on the UCL plateau and an equal-altitude criteria on the BC slope). Figure 11. View largeDownload slide Interpolation of the ambient noise measurement results along the cross-section I–I΄ (see Fig. 4) as a function of distance (x-axis) versus frequency (y-axis): (a) HVSR(f) amplitude; (b) ellipticity of the particle motion; (c) frequency of occurrence of polarized frequency; (d) cross-section I–I΄: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture; (6) seismic noise measurement station located just along the cross-section; (7) seismic noise measurement station projected on the cross-section (using a 20 m buffer on the UCL plateau and an equal-altitude criteria on the BC slope). Fig. 11 shows three contour plots where the x-axis represents the distance along the cross-section and the y-axis corresponds to frequency. The colour scale represents respectively: the amplitude of the HVSR(f) function (Fig. 11a), the ellipticity of the particle motion (Fig. 11b), the frequency of occurrence of the polarized frequency for every azimuthal direction (Fig. 11c). These graphs illustrate the different seismic response of the zones distinguished in the Selmun Promontory. The non-polarized fundamental frequency of the UCL plateau is at 1.5–2.0 Hz and the amplitude of the HVSR(f) function tends to increase on moving from the inner plateau towards the unstable plateau edge, where the strongly polarized particle motion at 3.3–3.5 Hz results in correspondence of the main joints (zone B). On the BC slope (zone E), the value of the non-polarized resonance frequency increases with the decreasing thickness of the BC layer. Evidence that seismic energy concentrated in a specific range of frequency (typically higher than 3.0 Hz) can be related to the vibrational behaviour of unstable rock blocks has been shown in previous studies (Got et al. 2010; Galea et al. 2014). In particular, the seismic response of a rock block is related to its eigenmodes, which depend on both the geometrical and mechanical properties of the rock block. As suggested by Got et al. (2010), the order of magnitude of the flexion mode of a rock block can be obtained by the Blevins (2001) eq. (2):   $${f_{ij}} = \frac{{\pi e}}{2}\ \left[ {{{\left( {\frac{i}{a}} \right)}^2} + {{\left( {\frac{j}{b}} \right)}^2}} \right]\sqrt {\frac{E}{{12\rho \left( {1 - {\nu ^2}} \right)}}} ,$$ (2)where i and j are the numbers of vibration antinodes (equal to 1 for the first resonance frequency), E the Young modulus, ν the Poisson ratio, ρ the density and a, b and e respectively the z (thickness), x (length, dimension parallel to the rear fracture), and y (width, dimension perpendicular to the rear fracture) dimensions of the rock block. The ρ value was derived by the laboratory tests (i.e. 2146 kg m−3). A standard value of 0.25 was assumed for ν, while E resulted equal to 1.1 × 109 by the following eq. (3):   $$E\ = \frac{5}{6}\ \ \rho \ {V_P}^2,$$ (3)where VP is the P-wave velocity value, that can be assessed equal to 800 m s−1 for UCL based on the available literature (Farrugia et al. 2016, 2017). Both values of ν and E are in a range typical for a weak rock considering the presence of vertical and persistent joints in zone B. As a matter of fact, even if UCL can be assumed as rock with medium strength according to the Deere & Miller (1966) classification, in the UCL plateau edge the presence of the persistent and dense joint net places this rock mass in a weaker rock category (Sitharam et al. 2001; Ramamurthy 2004). The values of dimensions a, b and e for the whole zone B were assumed respectively equal to 30 m, 100 m and 8.5 m by using the engineering geological model (Figs 4 and 5). The whole thickness of the UCL layer was assumed for the a value, by considering the joints persistent down to the top of the BC layer. On the other hand, the values of b and e were assumed by considering the joint distribution: the first was assumed based on the length of the rear fracture of zone B, while the second by considering the mean width of zone B perpendicularly to the rear fracture. In this way, a value of about 3.5 Hz was predicted for the first resonance frequency, in a very good agreement with the experimental value measured at Selmun Promontory, in zone B, by the seismic noise analysis. In the light of the obtained results, despite a dense net of open fractures, zone B behaves as a unique unstable rock block with a uniform seismic response. Therefore, the maximum expected landslide event can be related to the movement of the whole zone, that is, an event having higher intensity than the detachment of a single rock block from the UCL plateau edge. A reliable characterization of the resonance eigenmodes of an unstable rock block can be an important technical constraint which may be used to propose defence strategies for managing the landslide risk, for example in the case of seismic shaking. In recent years, several published works demonstrated that seismic shaking can be considered responsible for triggering or reactivating of gravity-induced slope instabilities at different scales (Havenith et al. 2002, 2003a,b; Bozzano et al. 2004, 2008, 2011; Bourdeau & Havenith 2008; Lenti et al. 2015; Martino et al. 2017). At the Selmun Promontory, the damage to the Għajn Ħadid Tower occurred mainly on October 12th 1856 during the seismic shaking related to the Crete earthquake (MW 7.7) that caused damage to buildings all along the Maltese Archipelago (Galea 2007). In the light of this, it is possible to conjecture that the damage of the tower was due to the interaction between the seismic waves and the polarized resonance frequency of the rock blocks at zone B, which favoured displacement along pre-existent joints and/or formation of new joints, and facilitated the masonry collapse (Fig. 12). Figure 12. View largeDownload slide Photograph showing two joints that are supposed to pass under the Għajn Ħadid Tower (red solid lines) and two large cracks crossing the masonry (blue dashed lines) related to its collapse; the arrow indicates the direction of the polarization in zone B. Figure 12. View largeDownload slide Photograph showing two joints that are supposed to pass under the Għajn Ħadid Tower (red solid lines) and two large cracks crossing the masonry (blue dashed lines) related to its collapse; the arrow indicates the direction of the polarization in zone B. 7 FUTURE PERSPECTIVES In the framework of strategies devoted to ensure safety of tourists as well as cultural heritage preservation, the above results can provide a basis for designing a monitoring network for early warning. In this regard, it could be relevant to plan a future permanent installation of a sensor network aimed at identifying incipient failures that can be precursors for the collapse of some portions of the cliff slope. Such an approach can be borrowed from earthquake engineering, where the monitoring of the main resonance frequencies of buildings and civil structures has already been experimentally implemented for detecting changes in stiffness during progressive damage (Doebling et al. 1996; Kim & Stubbs 2003; Clinton et al. 2006). By applying a monitoring approach based on analysis of continuous seismic noise measurements, at the test site of Chamousset (Vercors Massif, French Alps), Got et al. (2010) and Lévy et al. (2010, 2011) detected significant changes in the resonance frequency of a rock block approaching collapse, that is, a drop in the value of the resonance frequency (Lévy et al. 2010) and a decrease in the amplitude of the base noise level (Got et al. 2010). A similar analysis of continuous records of seismic ambient noise can be performed at the Selmun Promontory to capture anomalies of specific noise parameters in the frequency range of the eigenmode frequency of the unstable block (e.g. for zone B) that could be related to changes in the vibrational behaviour of the rock block and/or to an impending general collapse. Such a promising monitoring approach is still under study in the scientific community, aiming to identify and calibrate the parameters that could help to evaluate an aggravation of the stability conditions and an incipient failure of an unstable mass. An important issue is the need to filter these parameters with respect to the environmental conditions. In this regard, Valentin et al. (2017) observed a high sensitivity of the ratio amplitude parameters of the first eigenmode frequency of an unstable rock block with respect to wind conditions at the test site of Les Arches (Vercors Massif, French Alps). On the other hand, Bottelin et al. (2017) observed a variation of the amplitude peaks of the fundamental frequencies of a reinforced rock block related to temperature variations at the La Bourne test site (Vercors Massif, French Alps). The issues raised by such experiments show the importance of test sites at real scale of unstable rock slope and, in this perspective, the Selmun Promontory can represent a significant site where a similar monitoring approach can be tested starting from the consistent results obtained here. 8 CONCLUSIONS The cliff slope of the Selmun Promontory (Malta, Central Mediterranean Sea) is involved in an ongoing lateral spreading landslide process which threatens the touristic and cultural heritage site of Għajn Ħadid. This process originates from the geological setting of the area where a grained limestone (UCL) overlies a plastic clay (BC). The Selmun Promontory lateral spreading is associated with falls, slides and/or topples of different-size rock blocks. On 2015 and 2016, field investigations were carried out to obtain an engineering geological model of the Selmun Promontory by surveying the dense joint net of the outcropping limestone. Ambient seismic noise measurements were carried out to evaluate the vibrational behaviour as well as the seismic response of the area, moving from the inland limestone plateau toward the unstable limestone plateau edge and descending along the clayey slope down to the sea. The seismic data were analysed using both the HVSR method and polarization analysis. The results show a non-polarized 1-D resonance frequency on the top of the limestone plateau (1.5–1.8 Hz), an increase of the HVSR(f) function amplitude at frequencies characterized by polarized and linear particle motion in the unstable plateau edge, and a frequency-increasing non-polarized 1-D resonance descending the clayey slope down to the sea. This seismic response is strongly constrained by the engineering geological model of the Selmun Promontory since the results show that the strongest polarization exists in the unstable plateau edge zone, normally oriented with respect to the joint strike, and is more intense where the very large blocks are dislodged by opened joints. Close to the Għajn Ħadid Tower, the polarization analysis exhibited intensely polarized frequencies in a very close range of values ranging from 3.3 and 3.5 Hz, which can be quantitatively equated to the first eigenmode of the rock block considering its dimension and its geomechanical features. Given its uniform seismic response, such a zone can be considered as a unique unstable rock block despite a dense net of open fractures and the maximum expected landslide event can be related to the movement of the whole zone. In this regard, the damage of the Għajn Ħadid Tower, which occurred on October 12th 1856 during an earthquake, may have been induced by the interaction between seismic waves and the seismic response of this unstable zone. This interaction could have produced a preferential displacement along the joints, that facilitated the breaking of the masonry. On the other hand, the inland zone of the Selmun Promontory plateau, which is characterized by the absence of fractures and unstable blocks, does not show any polarization effects. These results demonstrate that noise measurements (i) are a suitable tool for detecting unstable zones in a landslide area involving densely jointed rock masses and (ii) are reliable for hazard zonation based on different stability levels characterized by specific seismic response of the unstable rock blocks (i.e. eigenmode frequency and evidence of polarization effects). The results obtained in this work should encourage the installation of a permanent sensor network at the Selmun Promontory for testing a monitoring system that may help in the design of intervention strategies for cultural heritage protection and to ensure the safety of visitors to the site. Acknowledgements The authors are grateful to Jan Burjánek for the use of the polarization analysis codes and to Mapping Unit, Malta Environment and Planning Authority for the provision of the digital topography of Malta. The authors also thank Marisa Regina and Gianmarco Rea for their useful support in field activities and in tests performed at Laboratory of Engineering Geology of the Department of Earth Sciences at ‘Sapienza’ University of Rome. Finally, the authors wish to thank the two anonymous reviewers whose useful and constructive suggestions made possible to improve the original manuscript. REFERENCES Bichler A., Bobrowsky P., Best M., Douma M., Hunter J., Calvert T., Burns R., 2004. Three-dimensional mapping of a landslide using a multi-geophysical approach: the Quesnel Forks landslide, Landslides , 1( 1), 29– 40. https://doi.org/10.1007/s10346-003-0008-7 Google Scholar CrossRef Search ADS   Blevins R.D., 2001. Formulas for Natural Frequencies and Mode Shape , Krieger Publishing. Bogoslovsky V.A., Ogilvy A.A., 1977. Geophysical methods for the investigation of landslides, Geophysics , 42( 3), 562– 571. https://doi.org/10.1190/1.1440727 Google Scholar CrossRef Search ADS   Bonnefoy-Claudet S., Cotton F., Bard P.Y., 2006. The nature of noise wavefield and its applications for site effects studies, Earth-Sci. Rev. , 79( 3–4), 205– 227. https://doi.org/10.1016/j.earscirev.2006.07.004 Google Scholar CrossRef Search ADS   Bottelin P. et al.  , 2013a. Spectral analysis of prone-to-fall rock compartments using ambient vibrations, J. Environ. Eng. Geophys. , 18( 4), 205– 217. https://doi.org/10.2113/JEEG18.4.205 Google Scholar CrossRef Search ADS   Bottelin P., Lévy C., Baillet L., Jongmans D., Gueguen P., 2013b. Modal and thermal analysis of Les Arches unstable rock column (Vercors massif, French Alps), Geophys. J. Int. , 194( 2), 849– 858. https://doi.org/10.1093/gji/ggt046 Google Scholar CrossRef Search ADS   Bottelin P. et al.  , 2014. Seismic and mechanical studies of the artificially triggered rockfall at Mount Néron (French Alps, December 2011), Nat. Hazards Earth Syst. Sci. , 14( 12), 3175– 3193. https://doi.org/10.5194/nhess-14-3175-2014 Google Scholar CrossRef Search ADS   Bottelin P., Baillet L., Larose E., Jongmans D., Hantz D., Brenguier O., Cadet H., Helmstetter A., 2017. Monitoring rock reinforcement works with ambient vibrations: La Bourne case study (Vercors, France), Eng. Geol. , 226, 136– 145. https://doi.org/10.1016/j.enggeo.2017.06.002 Google Scholar CrossRef Search ADS   Bourdeau C., Havenith H.-B., 2008. Site effects modelling applied to the slope affected by the Suusamyr earthquake (Kyrgyzstan, 1992), Eng. Geol. , 97( 3–4), 126– 145. https://doi.org/10.1016/j.enggeo.2007.12.009 Google Scholar CrossRef Search ADS   Bozzano F., Martino S., Naso G., Prestininzi A., Romeo R.W., Scarascia Mugnozza G., 2004. The large Salcito landslide triggered by the 2002 Molise, Italy, Earthquake, Earthq. Spectra , 20( S1), S95– S105. https://doi.org/10.1193/1.1768539 Google Scholar CrossRef Search ADS   Bozzano F., Cardarelli E., Cercato M., Lenti L., Martino S., Paciello A., Scarascia Mugnozza G., 2008. Engineering-geology model of the seismically-induced Cerda landslide (Sicily, Italy), Boll. Geofis. Teor. Appl. , 49( 2), 205– 225. Bozzano F., Lenti L., Martino S., Paciello A., Scarascia Mugnozza G., 2011. Evidences of landslide earthquake triggering due to self-excitation process, Int. J. Earth Sci. , 100( 4), 861– 879. https://doi.org/10.1007/s00531-010-0514-5 Google Scholar CrossRef Search ADS   Burjánek J., Gassner-Stamm G., Poggi V., Moore J.R., Fäh D., 2010. Ambient vibration analysis of an unstable mountain slope, Geophys. J. Int. , 180( 2), 820– 828. https://doi.org/10.1111/j.1365-246X.2009.04451.x Google Scholar CrossRef Search ADS   Burjánek J., Moore J.R., Yugsi Molina F.X., Fäh D., 2012. Instrumental evidence of normal mode rock slope vibration, Geophys. J. Int. , 188( 2), 559– 569. https://doi.org/10.1111/j.1365-246X.2011.05272.x Google Scholar CrossRef Search ADS   Chambers J.E. et al.  , 2011. Three-dimensional geophysical anatomy of an active landslide in Lias Group mudrocks, Cleveland Basin, UK, Geomorphology , 125( 4), 472– 484. https://doi.org/10.1016/j.geomorph.2010.09.017 Google Scholar CrossRef Search ADS   Clinton J.F., Case Bradford S., Heaton T.H., Favela J., 2006. The observed wander of the natural frequencies in a structure, Bull. seism. Soc. Am. , 96( 1), 237– 257. https://doi.org/10.1785/0120050052 Google Scholar CrossRef Search ADS   Colombero C., Baillet L., Comina C., Jongmans D., Vinciguerra S., 2017. Characterization of the 3-D fracture setting of an unstable rock mass: From surface and seismic investigations to numerical modeling, J. geophys. Res. , 122, 6346– 6366. https://doi.org/10.1002/2017JB014111 Google Scholar CrossRef Search ADS   Deere D.U., Miller R.P., 1966. Engineering Classification and Index Properties for Intact Rock , Illinois Univ. At Urbana Dept. of Civil Engineering. Del Gaudio V., Wasowski J., Pierri P., Mascia U., Calcagnile G., 2000. Gravimetric study of a retrogressive landslide in southern Italy, Surv. Geophys. , 21( 4), 391– 406. https://doi.org/10.1023/A:1006793009054 Google Scholar CrossRef Search ADS   Del Gaudio V., 2017. Instantaneous polarization analysis of ambient noise recordings in site response investigations, Geophys. J. Int. , 210( 1), 443– 464. https://doi.org/10.1093/gji/ggx175 Google Scholar CrossRef Search ADS   Del Gaudio V., Coccia S., Wasowski J., Gallipoli M.R., Mucciarelli M., 2008. Detection of directivity in seismic site response from microtremor spectral analysis, Nat. Hazards Earth Syst. Sci. , 8, 751– 762. https://doi.org/10.5194/nhess-8-751-2008 Google Scholar CrossRef Search ADS   Della Seta M., Martino S., Scarascia Mugnozza G., 2013. Quaternary sea-level change and slope instability in coastal areas: Insights from the Vasto Landslide (Adriatic coast, central Italy), Geomorphology , 201, 462– 478. https://doi.org/10.1016/j.geomorph.2013.07.019 Google Scholar CrossRef Search ADS   Dewez T.J.B., Rohmer J., Regard V., Cnudde C., 2013. Probabilistic coastal cliff collapse hazard from repeated terrestrial laser surveys: Case study from Mesnil Val (Normandy, northern France), J. Coast. Res. , Spec. Issue, 65, 702– 707. https://doi.org/10.2112/SI65-119.1 Google Scholar CrossRef Search ADS   Dickson M.E., Ogawa H., Kench P.S., Hutchinson A., 2013. Sea-cliff retreat and shore platform widening: steady-state equilibrium?, Earth Surf. Process. Landf. , 38( 9), 1046– 1048. https://doi.org/10.1002/esp.3422 Google Scholar CrossRef Search ADS   Doebling S.W., Farrar C.R., Prime M.B., Shevitz D.W., 1996. Damage Identification and Health Monitoring of Structural and Mechanical Systems from Changes in Their Vibration: A Literature Review. Technical Report LA-13070-MS , Los Alamos National Laboratory. Douze E.J., 1964. Rayleigh waves in short-period seismic noise, Bull. seism. Soc. Am. , 54( 4), 1197– 1212. https://doi.org/https://pubs.geoscienceworld.org/ssa/bssa/article/57/1/55/116458/short-period-seismic-noise Douze E.J., 1967. Short-period seismic noise, Bull. seism. Soc. Am. , 57( 1), 55– 81. Epifanio B., Zezere J.L., Neves M., 2013. Identification of hazardous zones combining cliff retreat rates with landslide susceptibility assessment, J. Coast. Res. , 65, 1681– 1686. https://doi.org/10.2112/SI65-284.1 Google Scholar CrossRef Search ADS   Farrugia D., Paolucci E., D’Amico S., Galea P., 2016. Inversion of surface wave data for subsurface shear wave velocity profiles characterized by a thick buried low-velocity layer, Geophys. J. Int. , 206( 2), 1221– 1231. https://doi.org/10.1093/gji/ggw204 Google Scholar CrossRef Search ADS   Farrugia D., Galea P., D’Amico S., Paolucci E., 2017. Sensitivity of ground motion parameters to local shear-wave velocity models: The case of buried low-velocity layers. Soil Dyn. Earthq. Eng. , 100, 196– 205. https://doi.org/10.1016/j.soildyn.2017.05.033 Google Scholar CrossRef Search ADS   Fiorucci M., Iannucci R., Lenti L., Martino S., Paciello A., Prestininzi A., Rivellino S., 2017. Nanoseismic monitoring of gravity-induced slope instabilities for the risk management of an aqueduct infrastructure in Central Apennines (Italy), Nat. Hazards , 86( S2), 345– 362. https://doi.org/10.1007/s11069-016-2516-5 Google Scholar CrossRef Search ADS   Galea P., 2007. Seismic history of the Maltese islands and considerations on seismic risk, Ann. Geophys. , 50( 6), 725– 740. https://doi.org/10.4401/ag-3053 Galea P., D’Amico S., Farrugia D., 2014. Dynamic characteristics of an active coastal spreading area using ambient noise measurements-Anchor Bay, Malta, Geophys. J. Int. , 199( 2), 1166– 1175. https://doi.org/10.1093/gji/ggu318 Google Scholar CrossRef Search ADS   Gigli G., Frodella W., Mugnai F., Tapete D., Cigna F., Fanti R., Intrieri E., Lombardi L., 2012. Instability mechanisms affecting cultural heritage sites in the Maltese Archipelago, Nat. Hazards Earth Syst. Sci. , 12( 6), 1883– 1903. https://doi.org/10.5194/nhess-12-1883-2012 Google Scholar CrossRef Search ADS   Glade T., Stark P., Dikau R., 2005. Determination of potential landslide shear plane depth using seismic refraction-a case study in Rheinhessen, Germany, Bull. Eng. Geol. Environ. , 64( 2), 151– 158. https://doi.org/10.1007/s10064-004-0258-1 Google Scholar CrossRef Search ADS   Got J.-L., Mourot P., Grangeon J., 2010. Pre-failure behaviour of an unstable limestone cliff from displacement and seismic data, Nat. Hazards Earth Syst. Sci. , 10( 4), 819– 829. https://doi.org/10.5194/nhess-10-819-2010 Google Scholar CrossRef Search ADS   Goudie A., 2004. Encyclopedia of Geomorphology , Routledge. Guzzetti F., Carrara A., Cardinali M., Reichenbach P., 1999. Landslide hazard evaluation: a review of current techniques and their application in a multi-scale study, Central Italy, Geomorphology , 31( 1-4), 181– 216. https://doi.org/10.1016/S0169-555X(99)00078-1 Google Scholar CrossRef Search ADS   Hack R., 2000. Geophysics for slope stability, Surv. Geophys. , 21( 4), 423– 448. https://doi.org/10.1023/A:1006797126800 Google Scholar CrossRef Search ADS   Hampton M.A., Griggs G.B., Edil T.B., Guy D.E., Kelley J.T., Komar P.D., Mickelson D.M., Shipman H.M., 2004. Processes that Govern the Formation and Evolution of Coastal Cliffs, in Formation, Evolution, and Stability of Coastal Cliffs-Status and Trends , pp. 1– 4, eds Hampton M.A., Griggs G.B., USGS - Professional Paper 1693. Havenith H.-B., Jongmans D., Abdrakmatov K., Trefois P., Delvaux D., Torgoev A., 2000. Geophysical investigations on seismically induced surface effects, case study of a landslide in the Suusamyr valley, Kyrgyzstan, Surv. Geophys. , 21, 349– 369. Havenith H.-B., Jongmans D., Faccioli E., Abdrakhmatov K., Bard P.-Y., 2002. Site effect analysis around the seismically induced Ananevo rockslide, Kyrgyzstan, Bull. seism. Soc. Am. , 92( 8), 3190– 3209. https://doi.org/10.1785/0120010206 Google Scholar CrossRef Search ADS   Havenith H.-B., Strom A., Jongmans D., Abdrakhmatov A., Delvaux D., Tréfois P., 2003a. Seismic triggering of landslides, Part A: Field evidence from the Northern Tien Shan, Nat. Hazards Earth Syst. Sci. , 3( 1/2), 135– 149. https://doi.org/10.5194/nhess-3-135-2003 Google Scholar CrossRef Search ADS   Havenith H.-B., Vanini M., Jongmans D., Faccioli E., 2003. Initiation of earthquake-induced slope failure: influence of topographical and other site specific amplification effects, J. Seismol. , 7( 3), 397– 412. https://doi.org/10.1023/A:1024534105559 Google Scholar CrossRef Search ADS   Hibert C., Grandjean G., Bitri A., Travelletti J., Malet J.-P., 2012. Characterizing landslides through geophysical data fusion: Example of the La Valette landslide (France), Eng. Geol. , 128, 23– 29. https://doi.org/10.1016/j.enggeo.2011.05.001 Google Scholar CrossRef Search ADS   Hoek E., Bray J.W., 1981. Rock Slope Engineering , Institution of Mining and Metallurgy. Horike M., 1985. Inversion of phase velocity of long-period microtremors to the S-wave-velocity structure down to the basement in urbanized areas, J. Phys. Earth , 33( 2), 59– 96. https://doi.org/10.4294/jpe1952.33.59 Google Scholar CrossRef Search ADS   Hutchinson J.N., 1988. General Report: Morphological and geotechnical parameters of landslides in relation to geology and hydrogeology, in Proceedings of Fifth International Symposium on Landslides , vol. 1, pp. 3– 35, ed. Bonnard C., Balkema. Google Scholar CrossRef Search ADS   Hutchinson J.N., Bromhead E.N., Chandler M.P., 1991. Investigations of landslides at St. Catherine's Point, Isle of Wight, in Slope Stability Engineering: Developments and Applications , pp. 169– 178, ed. Chandler M.P., Thomas Telford Publishing. Google Scholar CrossRef Search ADS   Hyde H.P.T., 1955. Geology of the Maltese Islands , Lux Press. Iannucci R., Martino S., Paciello A., D’Amico S., 2017. Rock mass characterization coupled with seismic noise measurements to analyze the unstable cliff slope of the Selmun Promontory (Malta), Procedia Eng. , 191, 263– 269. https://doi.org/10.1016/j.proeng.2017.05.180 Google Scholar CrossRef Search ADS   Imposa S., Grassi S., Fazio F., Rannisi G., Cino P., 2017. Geophysical surveys to study a landslide body (north-eastern Sicily), Nat. Hazards , 86( S2), 327– 343. https://doi.org/10.1007/s11069-016-2544-1 Google Scholar CrossRef Search ADS   ISRM, 1978. Suggested methods for the quantitative description of discontinuities in rock masses, Int. J. Rock Mech. Min. Sci. Geomech. Abstr. , 15, 319– 368. https://doi.org/10.1016/0148-9062(78)91472-9 CrossRef Search ADS   ISRM, 1979. Suggested method for determining water content, porosity, density, absorption and related properties and swelling and slake durability index properties, Int. J. Rock Mech. Min. Sci. Geomech. Abstr. , 16, 141– 156. Jongmans D., Garambois S., 2007. Geophysical investigation of landslides: A review, Bull. Soc. Geol. France , 178( 2), 101– 112. https://doi.org/10.2113/gssgfbull.178.2.101 Google Scholar CrossRef Search ADS   Kim J.T., Stubbs N., 2003. Crack detection in beam-type structures using frequency data, J. Sound Vib. , 259( 1), 145– 160. https://doi.org/10.1006/jsvi.2002.5132 Google Scholar CrossRef Search ADS   Konno K., Ohmachi T., 1998. Ground-motion characteristics estimated from spectral ratio between horizontal and vertical components of microtremor, Bull. seism. Soc. Am. , 88, 228– 241. Lenti L., Martino S., Paciello A., Prestininzi A., Rivellino S., 2012. Microseismicity within a karstified rock mass due to cracks and collapses as a tool for risk management, Nat. Hazards , 64( 1), 359– 379. https://doi.org/10.1007/s11069-012-0245-y Google Scholar CrossRef Search ADS   Lenti L., Martino S., Paciello A., Prestininzi A., Rivellino S., 2015. Recorded displacements in a landslide slope due to regional and teleseismic earthquakes, Geophys. J. Int. , 201( 3), 1335– 1345. https://doi.org/10.1093/gji/ggv063 Google Scholar CrossRef Search ADS   Lévy C., Baillet L., Jongmans D., Mourot P., Hantz D., 2010. Dynamic response of the Chamousset rock column (Western Alps, France), J. geophys. Res. , 115( F4), F04043, doi:10.1029/2009JF001606. https://doi.org/10.1029/2009JF001606 Google Scholar CrossRef Search ADS   Lévy C., Jongmans D., Baillet L., 2011. Analysis of seismic signals recorded on a prone-to-fall rock column (Vercors massif, French Alps), Geophys. J. Int. , 186( 1), 296– 310. https://doi.org/10.1111/j.1365-246X.2011.05046.x Google Scholar CrossRef Search ADS   Li T.M.C., Ferguson J.F., Herrin E., Durham H.B., 1984. High-frequency seismic noise at Lajitas, Texas, Bull. seism. Soc. Am. , 74( 5), 2015– 2033. Mantovani M., Devoto S., Forte E., Mocnik A., Pasuto A., Piacentini D., Soldati M., 2013. A multidisciplinary approach for rock spreading and block sliding investigation in the north-western coast of Malta, Landslides , 10( 5), 611– 622. https://doi.org/10.1007/s10346-012-0347-3 Google Scholar CrossRef Search ADS   Martino S., Mazzanti P., 2014. Integrating geomechanical surveys and remote sensing for sea cliff slope stability analysis: the Mt. Pucci case study (Italy), Nat. Hazards Earth Syst. Sci. , 14( 4), 831– 848. https://doi.org/10.5194/nhess-14-831-2014 Google Scholar CrossRef Search ADS   Martino S. et al.  , 2017. Ground effects triggered by the 24th August 2016, Mw 6.0 Amatrice (Italy) earthquake: surveys and inventorying to update the CEDIT catalogue, Geogr. Fis. Din. Quat. , 40, 77– 95. Maurer H., Spillmann T., Heincke B., Hauck C., Loew S., Springman S.M., Green A.G., 2010. Geophysical characterization of slope instabilities, First Break , 28( 8), 53– 61. McCann D.M., Forster A., 1990. Reconnaissance geophysical methods in landslide investigations, Eng. Geol. , 29( 1), 59– 78. https://doi.org/10.1016/0013-7952(90)90082-C Google Scholar CrossRef Search ADS   Méric O., Garambois S., Jongmans D., Wathelet M., Chatelain J.L., Vengeon J.M., 2005. Application of geophysical methods for the investigation of the large gravitational mass movement of Séchilienne, France, Can. Geotech. J. , 42( 4), 1105– 1115. https://doi.org/10.1139/t05-034 Google Scholar CrossRef Search ADS   Montoya-Montes I., Rodríguez-Santalla I., Sánchez-García M.J., Alcántara-Carrió J., Martín-Velázquez S., Gómez-Ortiz D., Martín-Crespo T., 2012. Mapping of landslide susceptibility of coastal cliffs: The Mont-Roig del Camp case study, Geol. Acta , 10, 439– 455. Moore J.R., Gischig V., Burjánek J., Loew S., Fäh D., 2011. Site effects in unstable rock slopes: dynamic behavior of the Randa instability (Switzerland), Bull. seism. Soc. Am. , 101( 6), 3110– 3116. https://doi.org/10.1785/0120110127 Google Scholar CrossRef Search ADS   Moore J.R. et al.  , 2016. Anthropogenic sources stimulate resonance of a natural rock bridge, Geophys. Res. Lett. , 43( 18), 9669– 9676. https://doi.org/10.1002/2016GL070088 Google Scholar CrossRef Search ADS   Nakamura Y., 1989. A method for dynamic characteristics estimation of subsurface using microtremor on the ground surface, Q. Rep. Railw. Tech. Res. Inst. , 30( 1), 25– 33. Nogoshi M., Igarashi T., 1970. On the propagation characteristics of microtremors, J. Seismol. Soc. Japan , 23, 264– 280. (in Japanese with English abstract). Nogoshi M., Igarashi T., 1971. On the amplitude characteristics of microtremors, J. Seismol. Soc. Japan , 24, 24– 40. (in Japanese with English abstract). Panzera F., D’Amico S., Lotteri A., Galea P., Lombardo G., 2012. Seismic site response of unstable steep slope using noise measurements: the case study of Xemxija Bay area, Malta, Nat. Hazards Earth Syst. Sci. , 12( 11), 3421– 3431. https://doi.org/10.5194/nhess-12-3421-2012 Google Scholar CrossRef Search ADS   Panzera F., D’Amico S., Galea P., Lombardo G., Gallipoli M.R., Pace S., 2013. Geophysical measurements for site response investigation: preliminary results on the island of Malta, Boll. Geofis. Teor. Appl. , 54, 111– 128. Panzera F., Pischiutta M., Lombardo G., Monaco C., Rovelli A., 2014. Wavefield polarization in fault zones of the Western flank of Mt. Etna: Observations and fracture orientation modelling, Pure appl. Geophys. , 171( 11), 3083– 3097. https://doi.org/10.1007/s00024-014-0831-x Google Scholar CrossRef Search ADS   Pazzi V., Tanteri L., Bicocchi G, D’Ambrosio M., Caselli A., Fanti R., 2017. H/V measurements as an effective tool for the reliable detection of landslide slip surfaces: Case studies of Castagnola (La Spezia, Italy) and Roccalbegna (Grosseto, Italy), Phys. Chem. Earth, Parts A/B/C , 98, 136– 153. https://doi.org/10.1016/j.pce.2016.10.014 Google Scholar CrossRef Search ADS   Pedley M., 2011. The Calabrian Stage, Pleistocene highstand in Malta: a new marker for unravelling the Late Neogene and Quaternary history of the islands, J. Geol. Soc. , 168( 4), 913– 926. https://doi.org/10.1144/0016-76492010-080 Google Scholar CrossRef Search ADS   Pedley H.M., House M.R., Waugh B., 1976. The geology of Malta and Gozo, Proc. Geologists' Assoc. , 87( 3), 325– 341. https://doi.org/10.1016/S0016-7878(76)80005-3 Google Scholar CrossRef Search ADS   Pedley H.M., House M.R., Waugh B., 1978. The geology of the Pelagian block: the Maltese islands, in The Ocean Basins and Margins , pp. 417– 433, eds Nairn A.E.M., Kanes W.H., Stehli F.G., Plenum Press. Google Scholar CrossRef Search ADS   Pedley M., Clarke M., Galea P., 2002. Limestone Isles in a Crystal Sea: The Geology of the Maltese Islands , Publishers Enterprises Group Ltd. Pennetta M., Russo E.L., 2011. Hazard factors in high rocky coasts of Capri Island (Gulf of Naples, Italy), J. Coast. Res. , 61, 428– 434. https://doi.org/10.2112/SI61-001.53 Google Scholar CrossRef Search ADS   Petley D., 2012. Global patterns of loss of life from landslides, Geology , 40( 10), 927– 930. https://doi.org/10.1130/G33217.1 Google Scholar CrossRef Search ADS   Pischiutta M. et al.  , 2017. Results from shallow geophysical investigations in the northwestern sector of the island of Malta, Phys. Chem. Earth, Parts A/B/C , 98, 41– 48. https://doi.org/10.1016/j.pce.2016.10.013 Google Scholar CrossRef Search ADS   Ramamurthy T., 2004. A geo-engineering classification for rocks and rock masses, Int. J. Rock Mech. Min. Sci. , 41( 1), 89– 101. https://doi.org/10.1016/S1365-1609(03)00078-9 Google Scholar CrossRef Search ADS   Rosser N.J., Brain M.J., Petley D.N., Lim M., Norman E.C., 2013. Coastline retreat via progressive failure of rocky coastal cliffs, Geology , 41( 8), 939– 942. https://doi.org/10.1130/G34371.1 Google Scholar CrossRef Search ADS   Sass O., Bell R., Glade T., 2008. Comparison of GPR, 2D-resistivity and traditional techniques for the subsurface exploration of the Öschingen landslide, Swabian Alb (Germany), Geomorphology , 93( 1-2), 89– 103. https://doi.org/10.1016/j.geomorph.2006.12.019 Google Scholar CrossRef Search ADS   Senfaute G., Duperret A., Lawrence J.A., 2009. Micro-seismic precursory cracks prior to rock-fall on coastal chalk cliffs: a case study at Mesnil-Val, Normandie, NW France, Nat. Hazards Earth Syst. Sci. , 9( 5), 1625– 1641. https://doi.org/10.5194/nhess-9-1625-2009 Google Scholar CrossRef Search ADS   Sitharam T.G., Sridevi J., Shimizu N., 2001. Practical equivalent continuum characterization of jointed rock masses, Int. J. Rock Mech. Min. Sci. , 38( 3), 437– 448. https://doi.org/10.1016/S1365-1609(01)00010-7 Google Scholar CrossRef Search ADS   Starr A.M., Moore J.R., Thorne M.S., 2015. Ambient resonance of Mesa Arch, Canyonlands National Park, Utah, Geophys. Res. Lett. , 42( 16), 6696– 6702. https://doi.org/10.1002/2015GL064917 Google Scholar CrossRef Search ADS   Tang C., Li L., Xu N., Ma K., 2015. Microseismic monitoring and numerical simulation on the stability of high-steep rock slopes in hydropower engineering, J. Rock Mech. Geotech. Eng. , 7( 5), 493– 508. https://doi.org/10.1016/j.jrmge.2015.06.010 Google Scholar CrossRef Search ADS   Toksöz M.N., Lacoss R.T., 1968. Microseisms: mode structure and sources, Science , 159( 3817), 872– 873. https://doi.org/10.1126/science.159.3817.872 Google Scholar CrossRef Search ADS PubMed  Valentin J., Capron A., Jongmans D., Baillet L., Bottelin P., Donze F., Larose E., Mangeney A., 2017. The dynamic response of prone-to-fall columns to ambient vibrations: comparison between measurements and numerical modelling, Geophys. J. Int. , 208( 2), 1058– 1076. https://doi.org/10.1093/gji/ggw440 Google Scholar CrossRef Search ADS   Varnes D.J., 1978. Slope movement types and processes, in Landslides Analysis and Control, Special Report 176 , pp. 11– 33, eds Schuster R.L., Krizek R.J., National Academy of Science. Varnes D.J. & IAEG, 1984. Landslide Hazard Zonation: A Review of Principles and Practice , United Nations Scientific and Cultural Organization. Vella A., Galea P., D’Amico S., 2013. Site frequency response characterisation of the Maltese islands based on ambient noise H/V ratios, Eng. Geol. , 163, 89– 100. https://doi.org/10.1016/j.enggeo.2013.06.006 Google Scholar CrossRef Search ADS   Vidale J.E., 1986. Complex polarisation analysis of particle motion, Bull. seism. Soc. Am. , 76, 1393– 1405. Xu N.W., Li T.B., Dai F., Li B., Zhu Y.G., Yang D.S., 2015. Microseismic monitoring and stability evaluation for the large scale underground caverns at the Houziyan hydropower station in Southwest China, Eng. Geol. , 188, 48– 67. https://doi.org/10.1016/j.enggeo.2015.01.020 Google Scholar CrossRef Search ADS   Yalcinkaya E. et al.  , 2016. Near-surface geophysical methods for investigating the Buyukcekmece landslide in Istanbul, Turkey, J. Appl. Geophys. , 134, 23– 35. https://doi.org/10.1016/j.jappgeo.2016.08.012 Google Scholar CrossRef Search ADS   Yamanaka H., Takemura M., Ishida H., Niwa M., 1994. Characteristics of long-period microtremors and their applicability in exploration of deep sedimentary layers, Bull. seism. Soc. Am. , 84( 6), 1831– 1841. Young A.P., Ashford S.A., 2008. Instability investigation of cantilevered seacliffs, Earth Surf. Process. Landf. , 33( 11), 1661– 1677. https://doi.org/10.1002/esp.1636 Google Scholar CrossRef Search ADS   Zang A. et al.  , 2017. Hydraulic fracture monitoring in hard rock at 410 m depth with an advanced fluid-injection protocol and extensive sensor array, Geophys. J. Int. , 208( 2), 790– 813. https://doi.org/10.1093/gji/ggw430 Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of The Royal Astronomical Society. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Geophysical Journal International Oxford University Press

# Engineering geological zonation of a complex landslide system through seismic ambient noise measurements at the Selmun Promontory (Malta)

, Volume 213 (2) – May 1, 2018
16 pages

/lp/ou_press/engineering-geological-zonation-of-a-complex-landslide-system-through-OyEgzJgTz0
Publisher
The Royal Astronomical Society
ISSN
0956-540X
eISSN
1365-246X
D.O.I.
10.1093/gji/ggy025
Publisher site
See Article on Publisher Site

### Abstract

Summary The cliff slope of the Selmun Promontory, located in the Northern part of the island of Malta (Central Mediterranean Sea) close to the coastline, is involved in a landslide process as exhibited by the large block-size talus at its bottom. The landslide process is related to the geological succession outcropping in the Selmun area, characterized by the overposition of a grained limestone on a plastic clay, that induces a lateral spreading phenomenon associated with detachment and collapse of different-size rock blocks. The landslide process shapes a typical landscape with a stable plateau of stiff limestone bordered by an unstable cliff slope. The ruins of Għajn Ħadid Tower, the first of the 13 watchtowers built in 1658 by the Grand Master Martin de Redin, stand out on the Selmun Promontory. The conservation of this important heritage site, already damaged by an earthquake which struck the Maltese Archipelago on 1856 October 12, is currently threatened by a progressive retreat of the landslide process towards the inland plateau area. During 2015 and 2016, field surveys were carried out to derive an engineering geological model of the Selmun Promontory. After a high-resolution geomechanical survey, the spatial distribution of the joints affecting the limestone was obtained. At the same time, 116 single-station noise measurements were carried out to cover inland and edge of the limestone plateau as well as the slope where the clays outcrop. The obtained 1-hour time histories were analysed through the horizontal to vertical spectral ratio technique, as well as polarization and ellipticity analysis of particle motion to define the local seismic response in zones having different stability conditions, that is, related to the presence of unstable rock blocks characterized by different vibrational modes. The results obtained demonstrate the suitability of passive seismic geophysical techniques for zoning landslide hazard in case of rock slopes and prove the relevance of anisotropies in conditioning the polarization of vibrational modes for dislodged rock masses. Geomechanics, Geomorphology, Seismic noise, Site effects 1 INTRODUCTION Landslides are one of the major natural hazards, causing large human and economic losses worldwide (Petley 2012). In coastal areas, landslides and related risk have been widely studied in the scientific community due to the presence of populated settlements as well as touristic and cultural heritage sites (Guzzetti et al. 1999; Senfaute et al. 2009; Pennetta & Russo 2011; Montoya-Montes et al. 2012; Dewez et al. 2013; Epifanio et al. 2013; Martino & Mazzanti 2014). Commonly, sea cliffs are subjected to retreat during their complex evolution, governed by a combination of sea erosion and gravity-induced instability processes on coastal cliff slopes (Hampton et al. 2004; Young & Ashford 2008; Dickson et al. 2013; Rosser et al. 2013). Such processes may produce landslides extending from a single sea cliff up to several hundred meters of coast (Hutchinson et al. 1991; Della Seta et al. 2013). The island of Malta (Central Mediterranean Sea) presents several significant case studies regarding coastal gravity- and erosion-induced instability processes. The geological setting of the Western region of the Maltese Archipelago, that is, the over-position of stiff limestone (Upper Coralline Limestone, UCL) on plastic clay deposits (Blue Clay, BC), leads to large lateral spreading processes associated with falls, slides and/or topples of different-size rock blocks. These instabilities affect both inland areas, for example, the plateaus on which the cities of Mdina (Malta) and the Citadel of Victoria (Gozo) are built (Gigli et al. 2012), as well as sea cliffs all along the coastline, especially in the northwestern part of Malta and in Gozo (Mantovani et al. 2013; Galea et al. 2014). In recent years, multidisciplinary approaches have been developed to investigate landslides by coupling traditional engineering geological surveying (ISRM 1978) with geophysical techniques (Bogoslovsky & Ogilvy 1977; McCann & Forster 1990; Hack 2000; Jongmans & Garambois 2007; Maurer et al. 2010) that present advantages such as flexibility, efficiency, ease of instrument deployment and rapidity of data processing. In general, geophysical methods are applied to provide a 2-D or 3-D imaging of landslide mass (e.g. dimensions and geometry) and some basic features (e.g. water content or seismic wave velocity), as evidenced by published works in which different approaches were applied to investigate landslide, for example, electrical resistivity tomography (Havenith et al. 2000; Hibert et al. 2012), spontaneous potential (Chambers et al. 2011), ground penetrating radar (Bichler et al. 2004; Sass et al. 2008), gravimetry (Del Gaudio et al. 2000), electromagnetism (Méric et al. 2005), seismic reflection (Bichler et al. 2004), seismic refraction (Havenith et al. 2000; Glade et al. 2005), seismic tomography (Meric et al. 2005) and Multi-channel Analysis of Surface Waves (Imposa et al. 2017). Passive seismic surveys and, in particular, seismic ambient noise measurements have already been applied to investigate landslide-affected slopes (Del Gaudio et al. 2008; Burjánek et al. 2010, 2012; Yalcinkaya et al. 2016; Del Gaudio 2017; Pazzi et al. 2017), unstable blocks of rocky cliffs (Got et al. 2010; Lévy et al. 2010, 2011; Panzera et al. 2012; Bottelin et al. 2013a,b, 2014, 2017; Galea et al. 2014; Colombero et al. 2017) and natural rock arches (Starr et al. 2015; Moore et al. 2016) by different approaches, for example, horizontal to vertical spectral ratio (HVSR) analysis, f–k analysis, site to reference spectral ratios, polarization analysis, noise spectral and amplitude variations. With respect to other geophysical methods, the published applications of passive seismic approaches demonstrated their capability to take into account the time dimension, with the aim of observing the variation of specific effects or parameters with time. Such a capability can be applied to characterize the dynamic behaviour of a landslide in terms of resonance frequency of the unstable mass (Got et al. 2010; Lévy et al. 2010; Bottelin et al. 2013a,b, 2017; Valentin et al. 2017). According to the published results, a joint application of HVSR analysis and polarization analysis appears to yield reliable estimates of the dimension of the landslide body and to characterize its seismic response (Burjánek et al. 2010, 2012; Galea et al. 2014). Characterization of the vibrational behaviour of landslide affected slopes can be considered a useful tool to design strategies for managing landslide risk, for example, for planning and monitoring consolidation intervention (Bottelin et al. 2017) or designing monitoring systems as already experienced in the case of drainage plants (Lenti et al. 2012; Fiorucci et al. 2017), hydroelectric power stations (Tang et al. 2015; Xu et al. 2015) and geothermal heat exchangers (Zang et al. 2017). In the light of the perspectives demonstrated by seismic noise measurements, in this work we investigate the rock cliff slope of the Selmun Promontory, located on the northwestern coast of Malta and affected by an important landslide process. We apply a joint approach, including HVSR analysis and polarization analysis. Such an approach was applied to study the local seismic response and the vibrational behaviour of the cliff slope as well as to evaluate its usefulness as a tool in the framework of landslide hazard zonation, that is, the procedure for dividing the land surface into homogeneous areas, and ranking of these areas according to their degree of stability (Varnes & IAEG 1984). The Selmun Promontory was chosen as a case study also because the ruins of Għajn Ħadid Tower, the first of the thirteen watchtowers built in 1658 by the Grand Master Martin de Redin, stand out prominently in this area. Currently the conservation of this important heritage site, already damaged by an earthquake that struck the Maltese Archipelago on 1856 October 12, is threatened by the continuous deformations related to the spreading process which is evolving over the years from the coastline towards the inland plateau area. In 2015 and 2016, a detailed engineering geological survey was carried out to define the geological setting of the Selmun Promontory and to obtain the mechanical properties of the jointed rock mass. At the same time, 116 single-station seismic noise measurements were carried out over a wide area on both unstable and stable regions. This allowed the characterization of the local seismic response in terms of zones having different stability conditions, mainly related to the presence of already isolated rock blocks with respect to adjacent stable zones. The obtained results represent a fundamental contribution in the framework of designing protection strategies for managing the landslide process in the Selmun Promontory, that is, for ensuring the safety of tourists and preserving the heritage site of Għajn Ħadid Tower. 2 GEOLOGICAL AND GEOMORPHOLOGICAL SETTINGS The Maltese Archipelago is composed of three main islands (Malta, Gozo and Comino) located in the middle of the Mediterranean Sea, about 100 km South from Sicily and 290 km East from Tunisia. The islands represent the only currently emergent part of an extensive shallow-water shelf that extends from Eastern Sicily to the Malta Graben, an important part of the threshold separating the Western and Eastern Mediterranean basins (Pedley 2011). Limestones and clays of Oligocene and Miocene epochs form the Maltese islands. The sequence of the marine sedimentary rocks of the Maltese Archipelago (Fig. 1) is composed of five main geological formations (Hyde 1955; Pedley et al. 1976, 1978, 2002; Gigli et al. 2012) whose oldest formation is Lower Coralline Limestone (LCL) of Oligocene age (Chattian). Figure 1. View largeDownload slide Geology of the Maltese Archipelago: (a) geology map, position of the Maltese islands in the Mediterranean Sea and location of the Selmun Promontory in Malta (in the red frame); (b) sketch of the sedimentary sequence. Figure 1. View largeDownload slide Geology of the Maltese Archipelago: (a) geology map, position of the Maltese islands in the Mediterranean Sea and location of the Selmun Promontory in Malta (in the red frame); (b) sketch of the sedimentary sequence. Above this limestone, three formations are present and outcrop at the Selmun Promontory (Fig. 2a). Starting from the lowermost layer, these are: Globigerina Limestone (GL) formation, a soft yellowish fine-grained limestone of Lower Miocene age (Aquitanian-Langhian); BC formation, a soft pelagic blue or greenish grey marl and limey clay of Middle Miocene age (Serravallian); UCL formation, pale grey and orange fossiliferous coarse-grained limestone, of Upper Miocene in age (Tortonian-Messinian). Figure 2. View largeDownload slide Geological and geomorphological setting of the Selmun Promontory: (a) Photograph showing the UCL cliff slope, the UCL–BC and BC–GL stratigraphic geological contacts and the debris covering the BC slope; (b–d) Photographs showing the evolution over time of the landslide process, from the development of fractures in the UCL plateau (left) to the detachment of UCL rock blocks (right). Figure 2. View largeDownload slide Geological and geomorphological setting of the Selmun Promontory: (a) Photograph showing the UCL cliff slope, the UCL–BC and BC–GL stratigraphic geological contacts and the debris covering the BC slope; (b–d) Photographs showing the evolution over time of the landslide process, from the development of fractures in the UCL plateau (left) to the detachment of UCL rock blocks (right). Based on the geological setting, the island of Malta can be divided in two main sectors (Fig. 1a): the southeastern part of Malta where only the two oldest formations outcrop, shaping relatively flat landscapes, and the northwestern part of the island, where the full sedimentary sequence is generally conserved and the UCL and the BC mainly outcrop. Due to the outcropping geological formations, unstable cliff slopes typically forming the edges of a summit UCL plateau overtopping gentle slopes of the BC (between 30° and 45°) form the typical landscapes of the northwestern part of Malta (Gigli et al. 2012; Mantovani et al. 2013; Galea et al. 2014). The Selmun Promontory, located on the North-facing coast of the northwestern sector of Malta, was chosen as a case study because it is involved in a significant landslide process, as shown by the large block-size talus and the dense joint net clearly visible at the top of the stiff plateau (Fig. 2). In particular, the geological succession of the Selmun Promontory, characterized by the overposition of stiff rocks on a plastic deposit (i.e. the stiff UCL on the plastic BC), leads to a lateral spreading phenomenon (Goudie 2004): the horizontal deformation affecting the clayey materials, having viscoplastic behaviour, induces fracturing of the overlying stiff rock. Lateral spreading shapes a plateau of stiff rock bordered by jointed unstable cliffs, favouring the detachment of single rock blocks by typical gravity-induced instability mechanisms, that is, planar sliding, wedge sliding, toppling and falling (Hoek & Bray 1981). The resulting landslide process should be defined as complex type according to Varnes (1978) and Hutchinson (1988). During 2015 and 2016, field activities and satellite image studies were carried out to better characterize the geological features of the Selmun Promontory. The stratigraphic succession of the Selmun Promontory (Fig. 2a) is composed of 20–30 m of the UCL, 30 m of the BC and the GL at the base. All the formations show almost horizontal or slightly NE-dipping (<5°) strata. On the gentle slope, both the BC and the GL rarely outcrop because a significant slope debris deposit resulting from the UCL plateau evolution generally covers them. The observed slope debris is composed of UCL clasts having heterogeneous sizes (from centimetre-scale clasts up to metre-size blocks) embedded in weathered BC and residual material of limestone dissolution. The slope debris deposits reach several meters of thickness, especially in the part close to the UCL cliff wall where the largest-size rock blocks are located. By considering the Selmun Promontory morphology, it can be schematized as suggested by Martino & Mazzanti (2014) for coastal slopes as the final stage of evolution of a sea cliff, that is, a cliff no longer directly affected by sea erosion (except during exceptional storm events) that retreats primarily by gravity-induced instability processes; at this stage, a beach separates the old sea cliff from the current shoreline. 3 ENGINEERING GEOLOGICAL SURVEY Detailed engineering geological field surveys were carried out to define the geomechanical properties of the jointed rock mass composing the cliff slope of the Selmun Promontory (Iannucci et al. 2017). First of all, the joints recognized in the rock mass (on the plateau surface as well as on the cliff wall) were mapped and plotted on a satellite view combining information derived by a GPS device and field observations, such as direction and length of joint segments (with a decimetre resolution). Then, according to the ISRM standard (ISRM 1978), each joint was characterized by a number of features, such as attitude (dip direction and dip) that allowed to define two main systems of sub-vertical joints at the Selmun Promontory: J1 prevalent in the NW zone (mean dip direction 330°, strike 60°) and J2 prevalent in the SE one (mean dip direction 45°, strike 135°). The two joint systems have a dip direction approximately orthogonal to the plateau edge direction and, therefore, a strike similar to it (Fig. 3). A GIS geo-database was implemented to inventory all the collected field data. Additionally, 18 cubic samples of the UCL were analysed through laboratory tests to characterize the rock mass matrix and allowed us to obtain, through the hydrostatic weighing method (ISRM 1979), a density (ρ) value of 2146 kg m−3 with a standard deviation of 141 kg m−3. Table 1 summarizes the main geomechanical parameters obtained for the surveyed joints in the rock mass of the Selmun Promontory. Figure 3. View largeDownload slide Results of the geomechanical survey performed at the Selmun Promontory: (a) Stereographic projection (equal-angle lower-hemisphere) of joint poles surveyed and planes of the two main joint systems; (b) Satellite view showing the location of the two main joint systems. Figure 3. View largeDownload slide Results of the geomechanical survey performed at the Selmun Promontory: (a) Stereographic projection (equal-angle lower-hemisphere) of joint poles surveyed and planes of the two main joint systems; (b) Satellite view showing the location of the two main joint systems. Table 1. Summary of the main geomechanical properties of the rock mass joints derived from field activities at the Selmun Promontory. Rock mass joints  ID system  Dip direction and dip (°)  Opening (cm)  Vertical offset (cm)  Hydraulic conditions  JCS (MPa)  JRC  J1  330/89  0–105 average 22  0–40 average 12  No flow  41.0  8–10  J2  45/88  0–70 average 16  0–30 average 16  No flow  41.0  8–10  Rock mass joints  ID system  Dip direction and dip (°)  Opening (cm)  Vertical offset (cm)  Hydraulic conditions  JCS (MPa)  JRC  J1  330/89  0–105 average 22  0–40 average 12  No flow  41.0  8–10  J2  45/88  0–70 average 16  0–30 average 16  No flow  41.0  8–10  View Large Table 2. Summary of the ambient noise measurements carried out in the several zones over the Selmun area (see Fig. 4). ID zone  Area (m2)  Measurements (n)  Measurement density (n m−2)  A  27 000  35  0.0013  B  1000  50  0.05  C  600  10  0.017  D  450  6  0.014  E  100 000  15  0.00015  ID zone  Area (m2)  Measurements (n)  Measurement density (n m−2)  A  27 000  35  0.0013  B  1000  50  0.05  C  600  10  0.017  D  450  6  0.014  E  100 000  15  0.00015  View Large An engineering geological model of the Selmun Promontory was obtained by combining the collected geological and geomechanical data and by mapping the spatial distribution of the joint net (Fig. 4); from this model, two orthogonal cross-sections of the Selmun Promontory were obtained (Fig. 5). Figure 4. View largeDownload slide Satellite view showing the geological setting of the Selmun Promontory: (1) Upper Coralline Limestone (Tortonian-Messinian); (2) Blue Clay (Serravallian); (3) Globigerina Limestone (Aquitanian-Langhian); (4) debris slope deposit (Quaternary); 5) attitude of strata; 6) fracture (the dashed line represents an inferred fracture); 7) Għajn Ħadid Tower; 8) seismic noise measurement station; 9) cross-section (see Fig. 5). Elevation is given by contour lines (m a.s.l.). The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 4. View largeDownload slide Satellite view showing the geological setting of the Selmun Promontory: (1) Upper Coralline Limestone (Tortonian-Messinian); (2) Blue Clay (Serravallian); (3) Globigerina Limestone (Aquitanian-Langhian); (4) debris slope deposit (Quaternary); 5) attitude of strata; 6) fracture (the dashed line represents an inferred fracture); 7) Għajn Ħadid Tower; 8) seismic noise measurement station; 9) cross-section (see Fig. 5). Elevation is given by contour lines (m a.s.l.). The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 5. View largeDownload slide Cross-sections (see Fig. 4) of the Selmun Promontory: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) seismic noise measurement station along the cross-section. Figure 5. View largeDownload slide Cross-sections (see Fig. 4) of the Selmun Promontory: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) seismic noise measurement station along the cross-section. Based on the engineering geological features, five zones were distinguished in the area (Fig. 4): one zone located in the inland plateau (A), characterized by the absence of ground fractures and unstable blocks; three zones located at the edge of the plateau (B–D), where a high density of ground fractures and unstable blocks exists; one zone located on the gentle slope (E), where the BC outcrops even if covered by debris. 4 SEISMIC NOISE ACQUISITION AND ANALYSIS During the 2015–2016 field activities, several geophysical campaigns of seismic noise measurements were carried out on the Selmun Promontory. Seismic ambient noise was recorded at 116 stations over an area of approximately 0.1 km2. Most of the noise measurement stations (Fig. 4) were distributed to cover the unstable zones of the cliff slope (zones B–D) as well as the stable UCL plateau (zone A), while a few measurements were carried out on the BC gentle slope (zone E). Details of the measurement campaigns are shown in Table 2. A three-component seismometer was deployed at each station: 84 measurements were carried out using an LE-3D/5 s seismometer by Lennartz Electronic GmbH coupled with a REFTEK 130–01 datalogger, set to a sampling frequency of 250 Hz; 32 measurements were carried out using a Tromino three-component portable seismograph (www.tromino.eu), set to a 128 or 256 Hz sampling frequency. In both cases, the seismic noise was recorded for 1 hour. Seismic noise is composed of ambient vibrations of the ground produced by random and uncontrolled sources, natural or related to human activity, such as tides, sea waves striking the coasts, wind, trees, buildings, industrial machinery, road traffic, trains, human footsteps, etc. In general, authors who investigated the composition of seismic noise (Douze 1964, 1967; Toksöz & Lacoss 1968; Li et al. 1984; Horike 1985; Yamanaka et al. 1994; Bonnefoy-Claudet et al. 2006) assume that it consists mainly of surface waves. The HVSR analysis, proposed by Nogoshi & Igarashi (1970, 1971), was applied by Nakamura (1989) to seismic noise measurements to evaluate the fundamental frequency of a site (f0), especially where there is a marked impedance contrast due to the presence of low shear-wave velocity layers above the seismic bedrock. A resonance peak in the distribution of the HVSR values versus natural frequency can be interpreted both in terms of SH-wave resonance in soft surface layers as well as in terms of the ellipticity of particle motion when the ambient noise wave train is made up predominantly of surface waves (Bonnefoy-Claudet et al. 2006). The wavefield is expected to be a combination of both types and the HVSR curve contains information about the shear wave velocity profile in shallow sediments (Galea et al. 2014). In this study, the HVSR analysis was carried out by using the open source software Geopsy (www.geopsy.org). Each recorded time history was divided into nonoverlapping windows of 40 s and the fast Fourier transform (FFT) was computed for the three spatial components ( north–south, east–west and up–down) in the frequency range between 1.0 and 60.0 Hz, using a 5 per cent cosine taper to each window. The Fourier spectra of each window were smoothed by the Konno–Ohmachi function (Konno & Ohmachi 1998). By taking the logarithmic average of the two horizontal components and averaging over the windows, the amplitude spectra and the H/V spectral ratio were finally computed for each single record. Vidale (1986) introduced in the seismological community the concept of polarization of the particle motion and a method to analyse it based on principal component analysis of the coherency matrix (i.e. complex covariance matrix), which is computed from analytical signals of a three-component seismogram and without time averaging. Based on the hypothesis that the wavefield within an unstable rock mass is dominated by normal mode vibration rather than horizontal propagation of seismic waves, Burjánek et al. (2010, 2012) implemented such an analysis by adopting a Continuous Wavelet Transform (CWT) in order to carry out a time–frequency domain analysis on a time history. WAVEPOL package (Burjánek et al. 2012) allows to obtain a 3-D ellipse representing the particle motion at each time-frequency pair, in a visual output expressing the frequency dependence of ellipticity and polarization. The ellipticity of the particle motion is defined as the ratio between the semi-minor axis and the semi-major axis of the ellipse (i.e. 1 represents circular particle motion and 0 purely linear motion) for each frequency. Two angle values indicate the polarization of the particle motion at each frequency: a strike value, representing the azimuth of the semi-major axis projected to the horizontal plane from North, and a dip value, describing the dip angle of the semi-major axis from the horizontal plane. Such a polarization analysis provides quite robust results about directivity effects of the particle motion, overcoming the bias that could be introduced by the denominator spectrum in the HVSR calculation (Panzera et al. 2014). In particular, by combining the outputs from such an analysis, that is, observing for the same frequencies a high degree of linearity on the ellipticity diagram and an azimuthal direction on the polar strike plot, it is possible to demonstrate polarization effects of the particle motion. 5 RESULTS AND DISCUSSION 5.1 HVSR analysis The results obtained from the HVSR analysis were grouped according to the five zones previously defined in the Selmun area (Fig. 6), in order to distinguish the vibrational behaviour related to engineering geological features. Figure 6. View largeDownload slide HVSR(f) functions grouped for different zones of the Selmun Promontory (see Fig. 4): the red curves represent examples of obtained HVSR(f) among the 116 measures, the solid black line is the average HVSR(f) function obtained by all curves in each zone and the dashed black lines show the standard deviation of all curves in each zone. Figure 6. View largeDownload slide HVSR(f) functions grouped for different zones of the Selmun Promontory (see Fig. 4): the red curves represent examples of obtained HVSR(f) among the 116 measures, the solid black line is the average HVSR(f) function obtained by all curves in each zone and the dashed black lines show the standard deviation of all curves in each zone. All the stations located on the UCL plateau (i.e. both inland plateau and on the fractured plateau edge) show HVSR curves characterized by a ubiquitous resonance peak in a narrow frequency range of 1.5–2.0 Hz followed by a sharp dip of the spectral ratio. The HVSR peak at around 1.5–2.0 Hz can be considered representative of the fundamental frequency of the site (f0) since this peak is always present in the Maltese archipelago where the UCL–BC–GL sedimentary sequence is located, as evidenced by the studies carried out at Dingli, Mdina, Wardija, Xemxija, Mellieħa, Anchor Bay, Mgarr, Nadur, Rabat, Golden Bay and Bahrija (Panzera et al. 2012, 2013; Vella et al. 2013; Galea et al. 2014; Farrugia et al. 2016, 2017; Pischiutta et al. 2017). This HVSR feature (peak at 1.5–2.0 Hz followed by a dip below 1.0 over a wide frequency range) was interpreted as related to the difference of seismic wave velocity values between the three geological formations that compose the sedimentary sequence. The dip of the HVSR(f) function after 2.0 Hz is presumably due to the shallow shear-wave velocity inversion at the interface between the competent UCL and the plastic BC, while the peak can be related to Rayleigh wave ellipticity and/or trapping of SH waves in the BC low-velocity layer (Galea et al. 2014). On the other hand, the HVSR(f) functions on the UCL plateau show marked differences in the seismic response between the unstable zones and the stable plateau area at frequencies higher than 3.0 Hz (Fig. 6). As a matter of fact, in the measurements carried out within and in proximity of the unstable zones, the HVSR curves show significant additional resonance peaks at higher frequency (3.0–60.0 Hz) that are not present in the measurements carried out on the stable plateau zone. In addition, peaks of the HVSR(f) functions are much more evident in areas having higher density of fractures and blocks, such as the jointed edge of the UCL plateau. Zone B shows a prevalent HVSR peak at frequency around 3.5 Hz, while the other jointed edge zones (C and D) do not show peculiar characteristic HVSR peaks. The HVSR(f) functions resulting from the noise measurements on the BC slope show a first main peak in a range between 2.4 and 6.5 Hz, therefore at higher frequencies than the stations located on the UCL plateau. This peak can be related to the resonance in the BC layer due to the lower values of the seismic wave velocity compared to the underlying GL layer. A rough evaluation of the fundamental frequency of a resonant layer (f0) can be theoretically obtained by the following relation (1):   $${f_0} = \frac{{{V_S}}}{{4h}},$$ (1)where VS is the S-wave velocity within the layer and h its thickness. The BC formation is characterized by an S-wave velocity of 300–400 m s−1 (Panzera et al. 2012; Farrugia et al. 2016, 2017) and its maximum thickness at the Selmun Promontory is 30 m. The seismic noise measurements were carried out on the debris deposit, about 1–5 m thick. For the measurements carried out in the higher part of the slope, we can assume a total thickness of the BC–debris layer of 35 m and a mean S-wave velocity of 350 m s−1, by considering also the poorer geomechanical features of the debris. Such parameters yield a fundamental frequency (f0) of about 2.5 Hz in the higher part of the slope, which increases to 5 Hz by halving the thickness of the BC–debris layer, a condition existing in the middle part of the slope. The obtained values of fundamental frequency (f0) are similar to those obtained by Panzera et al. (2012) at several sites on Malta where the BC outcrops, that is, Rabat, Mdina, Wardija and Mellieħa. The HVSR(f) functions in this region show additional peaks at higher frequency that can be related to the resonance in the debris deposit due to its local features (i.e. thickness and/or weathering), since most of the measurements on the BC clay were carried out on the debris that covers the slope (Fig. 4). Fig. 7 summarizes the HVSR results for the surveyed area, showing the spatial distribution of the frequency values and the HVSR(f) function amplitudes of the first HVSR peak. On the UCL plateau (zones A-D), the frequency value of the first peak of the HVSR(f) function is between 1.5 and 1.8 Hz. The variation of the main peak amplitude can be related to the location of the measurement points: the HVSR(f) function amplitude increases on moving from the inner plateau towards the edge of the plateau, where the joint density significantly increases. As an example, the Southern portion of zone A is characterized by HVSR(f) amplitude values between 2.0 and 4.0, while in zone B the HVSR(f) amplitude ranges from 6.0 to 10.0. Figure 7. View largeDownload slide Satellite view showing the measurement points (dots) and the values (in white) of the first peak of the HVSR(f) function (colour scale): (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) Għajn Ħadid Tower. The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 7. View largeDownload slide Satellite view showing the measurement points (dots) and the values (in white) of the first peak of the HVSR(f) function (colour scale): (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture (the dashed line represents an inferred fracture); (6) Għajn Ħadid Tower. The light grey lines indicate the different zones of the Selmun Promontory (A–E). On the BC slope (zone E), the frequency of the first peak of the HVSR(f) function varies over a wider range compared to that on the UCL plateau, that is, between 2.4 and 6.5 Hz. Stations located in the higher part of the slope show lower resonance frequencies than the stations in the lower part. This variation can be explained by the decreasing thickness of BC on going down-slope, as shown in the engineering-geological cross-sections (Fig. 5). 5.2 Polarization analysis The polarization analysis confirms the peculiar features observed in the HVSR analysis and provides additional evidence to support the different seismic response in the various zones on the Selmun Promontory (Figs 8 and 9). Based on the polarization analysis, two features emerge with reference to specific frequency ranges: (i) the ellipse axes ratio, which decreases when the linearity of the particle motion increases; (ii) the azimuthal distribution of the particle motion, which reveals a main direction of polarization of the particle motion. Figure 8. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones A, B and C of the Selmun Promontory. Figure 8. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones A, B and C of the Selmun Promontory. Figure 9. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones D and E of the Selmun Promontory. Figure 9. View largeDownload slide Representative examples of HVSR(f) function (the dashed black lines show the standard deviation of the curve), ellipticity diagram and polar strike plot (same palette colour for relative frequency of occurrence) obtained for zones D and E of the Selmun Promontory. By analysing all the results obtained on both the UCL plateau and the BC slope, the frequencies associated with the first peak of the HVSR(f) function (i.e. 1.5–1.8 Hz for the UCL plateau and 2.4–6.5 Hz for the BC slope) do not show evidence of polarization. As a matter of fact, even if the ellipticity diagrams show quite a high degree of linearity (an ellipse axes ratio in a 0.1–0.2 range), the polar strike plots do not demonstrate a main direction of the particle motion at these frequencies. This confirms the hypothesis that the first peak of HVSR(f) function is mainly related to the stratigraphic setting (i.e. 1-D resonance effect) and to the multi-directional nature of the ambient noise. On the inland UCL plateau (i.e. zone A), the noise analysis shows no evidence of either peaks of the HVSR(f) function (beyond the f0 peak) or polarization effects of the particle motion, except for a few cases where these effects can be observed at frequency values higher than 30 Hz and are probably related to very local elements (e.g. thin and shallow soil layers characterized by low Vs values). Zone B shows strong directivity effects at the same frequency of the 1-D resonance which results from the HVSR analysis. The polarization analysis shows that in zone B the 3.3–3.5 Hz frequency range is characterized by marked features of linearity (ellipse axes ratio 0.0–0.1) as well as by a strong azimuthal polarization of the particle motion. As a matter of fact, the polar strike plots show a directivity of the particle motion at azimuthal values between 135° and 150°, values similar to the dip direction of the joint system prevalent in zone B (J1). This result is analogous to those obtained by Burjánek et al. (2010) and Moore et al. (2011) at the landslide of Randa (Switzerland) as well as by Galea et al. (2014) at the unstable sea cliff of Anchor Bay (Malta). In zone D, the polarization effects occur along the azimuthal direction 45–50°, values similar to the dip direction of the main joint system of the area (J2). Nevertheless, these effects occur at two specific frequency values: 5 Hz closest to the cliff (0–5 m from the edge) and 7 Hz in the internal portion of the plateau edge (5–10 m from the edge). There is also evidence of polarization effects along a direction similar to the joint dip direction in zone C but not in a specific range of frequency. Finally, in zone E, the analysis does not show polarization effects, despite the several peaks observed in the HVSR(f) functions. Fig. 10 maps the results obtained through the polarization analysis and shows the spatial distribution of the first polarized frequency values (only for frequencies lower than 30 Hz) at the Selmun Promontory, the corresponding azimuth and the associated ellipticity values. Such a spatial distribution confirms the absence of polarized particle motion on the stable plateau (zone A), while highlighting different behaviours in the unstable areas (zones B–D). In particular, even though zone B is characterized by a denser joint net, its uniform seismic response leads us to consider a vibrational behaviour related to a unique rock block characterized by a single polarized frequency between 3.3 and 3.5 Hz. On the other hand, in zone D the seismic response seems to be related to two different, or independently vibrating rock blocks, as shown by the two polarized frequency values (5 and 7 Hz). Figure 10. View largeDownload slide Satellite view summarizing the results of polarization analysis for the first (lower value) polarized frequency: (1) UCL; (2) debris slope deposit; (3) fracture (the dashed line represents an inferred fracture); (4) Għajn Ħadid Tower. Each symbol indicates a noise measurement point: the label displays the value of the first polarized frequency; the arrow is oriented in the azimuthal direction of the polarization; the colour scale indicates the ellipticity intensity. Black dots indicate noise measurement points without polarization effects. The light grey lines indicate the different zones of the Selmun Promontory (A–E). Figure 10. View largeDownload slide Satellite view summarizing the results of polarization analysis for the first (lower value) polarized frequency: (1) UCL; (2) debris slope deposit; (3) fracture (the dashed line represents an inferred fracture); (4) Għajn Ħadid Tower. Each symbol indicates a noise measurement point: the label displays the value of the first polarized frequency; the arrow is oriented in the azimuthal direction of the polarization; the colour scale indicates the ellipticity intensity. Black dots indicate noise measurement points without polarization effects. The light grey lines indicate the different zones of the Selmun Promontory (A–E). 6 ZONATION BASED ON VIBRATIONAL EFFECTS The different vibrational behaviour of the zones of the Selmun Promontory is shown in Fig. 11 where the results obtained for each noise station were projected and interpolated along the cross-section I–I’ (see Fig. 4 for location) with a gridding process using a triangulation with linear interpolation method. Additional seismic noise stations were projected on the cross-section to increase the density of results and improve the interpolation process: for the UCL plateau, we interpolated stations in a 20 m buffer from the track; for the BC slope we interpolated stations using an equal-altitude criteria, since the thicknesses variation of the BC layer and the debris deposit are negligible at the same altitude. Figure 11. View largeDownload slide Interpolation of the ambient noise measurement results along the cross-section I–I΄ (see Fig. 4) as a function of distance (x-axis) versus frequency (y-axis): (a) HVSR(f) amplitude; (b) ellipticity of the particle motion; (c) frequency of occurrence of polarized frequency; (d) cross-section I–I΄: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture; (6) seismic noise measurement station located just along the cross-section; (7) seismic noise measurement station projected on the cross-section (using a 20 m buffer on the UCL plateau and an equal-altitude criteria on the BC slope). Figure 11. View largeDownload slide Interpolation of the ambient noise measurement results along the cross-section I–I΄ (see Fig. 4) as a function of distance (x-axis) versus frequency (y-axis): (a) HVSR(f) amplitude; (b) ellipticity of the particle motion; (c) frequency of occurrence of polarized frequency; (d) cross-section I–I΄: (1) UCL; (2) BC; (3) GL; (4) debris slope deposit; (5) fracture; (6) seismic noise measurement station located just along the cross-section; (7) seismic noise measurement station projected on the cross-section (using a 20 m buffer on the UCL plateau and an equal-altitude criteria on the BC slope). Fig. 11 shows three contour plots where the x-axis represents the distance along the cross-section and the y-axis corresponds to frequency. The colour scale represents respectively: the amplitude of the HVSR(f) function (Fig. 11a), the ellipticity of the particle motion (Fig. 11b), the frequency of occurrence of the polarized frequency for every azimuthal direction (Fig. 11c). These graphs illustrate the different seismic response of the zones distinguished in the Selmun Promontory. The non-polarized fundamental frequency of the UCL plateau is at 1.5–2.0 Hz and the amplitude of the HVSR(f) function tends to increase on moving from the inner plateau towards the unstable plateau edge, where the strongly polarized particle motion at 3.3–3.5 Hz results in correspondence of the main joints (zone B). On the BC slope (zone E), the value of the non-polarized resonance frequency increases with the decreasing thickness of the BC layer. Evidence that seismic energy concentrated in a specific range of frequency (typically higher than 3.0 Hz) can be related to the vibrational behaviour of unstable rock blocks has been shown in previous studies (Got et al. 2010; Galea et al. 2014). In particular, the seismic response of a rock block is related to its eigenmodes, which depend on both the geometrical and mechanical properties of the rock block. As suggested by Got et al. (2010), the order of magnitude of the flexion mode of a rock block can be obtained by the Blevins (2001) eq. (2):   $${f_{ij}} = \frac{{\pi e}}{2}\ \left[ {{{\left( {\frac{i}{a}} \right)}^2} + {{\left( {\frac{j}{b}} \right)}^2}} \right]\sqrt {\frac{E}{{12\rho \left( {1 - {\nu ^2}} \right)}}} ,$$ (2)where i and j are the numbers of vibration antinodes (equal to 1 for the first resonance frequency), E the Young modulus, ν the Poisson ratio, ρ the density and a, b and e respectively the z (thickness), x (length, dimension parallel to the rear fracture), and y (width, dimension perpendicular to the rear fracture) dimensions of the rock block. The ρ value was derived by the laboratory tests (i.e. 2146 kg m−3). A standard value of 0.25 was assumed for ν, while E resulted equal to 1.1 × 109 by the following eq. (3):   $$E\ = \frac{5}{6}\ \ \rho \ {V_P}^2,$$ (3)where VP is the P-wave velocity value, that can be assessed equal to 800 m s−1 for UCL based on the available literature (Farrugia et al. 2016, 2017). Both values of ν and E are in a range typical for a weak rock considering the presence of vertical and persistent joints in zone B. As a matter of fact, even if UCL can be assumed as rock with medium strength according to the Deere & Miller (1966) classification, in the UCL plateau edge the presence of the persistent and dense joint net places this rock mass in a weaker rock category (Sitharam et al. 2001; Ramamurthy 2004). The values of dimensions a, b and e for the whole zone B were assumed respectively equal to 30 m, 100 m and 8.5 m by using the engineering geological model (Figs 4 and 5). The whole thickness of the UCL layer was assumed for the a value, by considering the joints persistent down to the top of the BC layer. On the other hand, the values of b and e were assumed by considering the joint distribution: the first was assumed based on the length of the rear fracture of zone B, while the second by considering the mean width of zone B perpendicularly to the rear fracture. In this way, a value of about 3.5 Hz was predicted for the first resonance frequency, in a very good agreement with the experimental value measured at Selmun Promontory, in zone B, by the seismic noise analysis. In the light of the obtained results, despite a dense net of open fractures, zone B behaves as a unique unstable rock block with a uniform seismic response. Therefore, the maximum expected landslide event can be related to the movement of the whole zone, that is, an event having higher intensity than the detachment of a single rock block from the UCL plateau edge. A reliable characterization of the resonance eigenmodes of an unstable rock block can be an important technical constraint which may be used to propose defence strategies for managing the landslide risk, for example in the case of seismic shaking. In recent years, several published works demonstrated that seismic shaking can be considered responsible for triggering or reactivating of gravity-induced slope instabilities at different scales (Havenith et al. 2002, 2003a,b; Bozzano et al. 2004, 2008, 2011; Bourdeau & Havenith 2008; Lenti et al. 2015; Martino et al. 2017). At the Selmun Promontory, the damage to the Għajn Ħadid Tower occurred mainly on October 12th 1856 during the seismic shaking related to the Crete earthquake (MW 7.7) that caused damage to buildings all along the Maltese Archipelago (Galea 2007). In the light of this, it is possible to conjecture that the damage of the tower was due to the interaction between the seismic waves and the polarized resonance frequency of the rock blocks at zone B, which favoured displacement along pre-existent joints and/or formation of new joints, and facilitated the masonry collapse (Fig. 12). Figure 12. View largeDownload slide Photograph showing two joints that are supposed to pass under the Għajn Ħadid Tower (red solid lines) and two large cracks crossing the masonry (blue dashed lines) related to its collapse; the arrow indicates the direction of the polarization in zone B. Figure 12. View largeDownload slide Photograph showing two joints that are supposed to pass under the Għajn Ħadid Tower (red solid lines) and two large cracks crossing the masonry (blue dashed lines) related to its collapse; the arrow indicates the direction of the polarization in zone B. 7 FUTURE PERSPECTIVES In the framework of strategies devoted to ensure safety of tourists as well as cultural heritage preservation, the above results can provide a basis for designing a monitoring network for early warning. In this regard, it could be relevant to plan a future permanent installation of a sensor network aimed at identifying incipient failures that can be precursors for the collapse of some portions of the cliff slope. Such an approach can be borrowed from earthquake engineering, where the monitoring of the main resonance frequencies of buildings and civil structures has already been experimentally implemented for detecting changes in stiffness during progressive damage (Doebling et al. 1996; Kim & Stubbs 2003; Clinton et al. 2006). By applying a monitoring approach based on analysis of continuous seismic noise measurements, at the test site of Chamousset (Vercors Massif, French Alps), Got et al. (2010) and Lévy et al. (2010, 2011) detected significant changes in the resonance frequency of a rock block approaching collapse, that is, a drop in the value of the resonance frequency (Lévy et al. 2010) and a decrease in the amplitude of the base noise level (Got et al. 2010). A similar analysis of continuous records of seismic ambient noise can be performed at the Selmun Promontory to capture anomalies of specific noise parameters in the frequency range of the eigenmode frequency of the unstable block (e.g. for zone B) that could be related to changes in the vibrational behaviour of the rock block and/or to an impending general collapse. Such a promising monitoring approach is still under study in the scientific community, aiming to identify and calibrate the parameters that could help to evaluate an aggravation of the stability conditions and an incipient failure of an unstable mass. An important issue is the need to filter these parameters with respect to the environmental conditions. In this regard, Valentin et al. (2017) observed a high sensitivity of the ratio amplitude parameters of the first eigenmode frequency of an unstable rock block with respect to wind conditions at the test site of Les Arches (Vercors Massif, French Alps). On the other hand, Bottelin et al. (2017) observed a variation of the amplitude peaks of the fundamental frequencies of a reinforced rock block related to temperature variations at the La Bourne test site (Vercors Massif, French Alps). The issues raised by such experiments show the importance of test sites at real scale of unstable rock slope and, in this perspective, the Selmun Promontory can represent a significant site where a similar monitoring approach can be tested starting from the consistent results obtained here. 8 CONCLUSIONS The cliff slope of the Selmun Promontory (Malta, Central Mediterranean Sea) is involved in an ongoing lateral spreading landslide process which threatens the touristic and cultural heritage site of Għajn Ħadid. This process originates from the geological setting of the area where a grained limestone (UCL) overlies a plastic clay (BC). The Selmun Promontory lateral spreading is associated with falls, slides and/or topples of different-size rock blocks. On 2015 and 2016, field investigations were carried out to obtain an engineering geological model of the Selmun Promontory by surveying the dense joint net of the outcropping limestone. Ambient seismic noise measurements were carried out to evaluate the vibrational behaviour as well as the seismic response of the area, moving from the inland limestone plateau toward the unstable limestone plateau edge and descending along the clayey slope down to the sea. The seismic data were analysed using both the HVSR method and polarization analysis. The results show a non-polarized 1-D resonance frequency on the top of the limestone plateau (1.5–1.8 Hz), an increase of the HVSR(f) function amplitude at frequencies characterized by polarized and linear particle motion in the unstable plateau edge, and a frequency-increasing non-polarized 1-D resonance descending the clayey slope down to the sea. This seismic response is strongly constrained by the engineering geological model of the Selmun Promontory since the results show that the strongest polarization exists in the unstable plateau edge zone, normally oriented with respect to the joint strike, and is more intense where the very large blocks are dislodged by opened joints. Close to the Għajn Ħadid Tower, the polarization analysis exhibited intensely polarized frequencies in a very close range of values ranging from 3.3 and 3.5 Hz, which can be quantitatively equated to the first eigenmode of the rock block considering its dimension and its geomechanical features. Given its uniform seismic response, such a zone can be considered as a unique unstable rock block despite a dense net of open fractures and the maximum expected landslide event can be related to the movement of the whole zone. In this regard, the damage of the Għajn Ħadid Tower, which occurred on October 12th 1856 during an earthquake, may have been induced by the interaction between seismic waves and the seismic response of this unstable zone. This interaction could have produced a preferential displacement along the joints, that facilitated the breaking of the masonry. On the other hand, the inland zone of the Selmun Promontory plateau, which is characterized by the absence of fractures and unstable blocks, does not show any polarization effects. These results demonstrate that noise measurements (i) are a suitable tool for detecting unstable zones in a landslide area involving densely jointed rock masses and (ii) are reliable for hazard zonation based on different stability levels characterized by specific seismic response of the unstable rock blocks (i.e. eigenmode frequency and evidence of polarization effects). The results obtained in this work should encourage the installation of a permanent sensor network at the Selmun Promontory for testing a monitoring system that may help in the design of intervention strategies for cultural heritage protection and to ensure the safety of visitors to the site. Acknowledgements The authors are grateful to Jan Burjánek for the use of the polarization analysis codes and to Mapping Unit, Malta Environment and Planning Authority for the provision of the digital topography of Malta. The authors also thank Marisa Regina and Gianmarco Rea for their useful support in field activities and in tests performed at Laboratory of Engineering Geology of the Department of Earth Sciences at ‘Sapienza’ University of Rome. Finally, the authors wish to thank the two anonymous reviewers whose useful and constructive suggestions made possible to improve the original manuscript. REFERENCES Bichler A., Bobrowsky P., Best M., Douma M., Hunter J., Calvert T., Burns R., 2004. Three-dimensional mapping of a landslide using a multi-geophysical approach: the Quesnel Forks landslide, Landslides , 1( 1), 29– 40. https://doi.org/10.1007/s10346-003-0008-7 Google Scholar CrossRef Search ADS   Blevins R.D., 2001. Formulas for Natural Frequencies and Mode Shape , Krieger Publishing. Bogoslovsky V.A., Ogilvy A.A., 1977. Geophysical methods for the investigation of landslides, Geophysics , 42( 3), 562– 571. https://doi.org/10.1190/1.1440727 Google Scholar CrossRef Search ADS   Bonnefoy-Claudet S., Cotton F., Bard P.Y., 2006. The nature of noise wavefield and its applications for site effects studies, Earth-Sci. Rev. , 79( 3–4), 205– 227. https://doi.org/10.1016/j.earscirev.2006.07.004 Google Scholar CrossRef Search ADS   Bottelin P. et al.  , 2013a. Spectral analysis of prone-to-fall rock compartments using ambient vibrations, J. Environ. Eng. Geophys. , 18( 4), 205– 217. https://doi.org/10.2113/JEEG18.4.205 Google Scholar CrossRef Search ADS   Bottelin P., Lévy C., Baillet L., Jongmans D., Gueguen P., 2013b. Modal and thermal analysis of Les Arches unstable rock column (Vercors massif, French Alps), Geophys. J. Int. , 194( 2), 849– 858. https://doi.org/10.1093/gji/ggt046 Google Scholar CrossRef Search ADS   Bottelin P. et al.  , 2014. Seismic and mechanical studies of the artificially triggered rockfall at Mount Néron (French Alps, December 2011), Nat. Hazards Earth Syst. Sci. , 14( 12), 3175– 3193. https://doi.org/10.5194/nhess-14-3175-2014 Google Scholar CrossRef Search ADS   Bottelin P., Baillet L., Larose E., Jongmans D., Hantz D., Brenguier O., Cadet H., Helmstetter A., 2017. Monitoring rock reinforcement works with ambient vibrations: La Bourne case study (Vercors, France), Eng. Geol. , 226, 136– 145. https://doi.org/10.1016/j.enggeo.2017.06.002 Google Scholar CrossRef Search ADS   Bourdeau C., Havenith H.-B., 2008. Site effects modelling applied to the slope affected by the Suusamyr earthquake (Kyrgyzstan, 1992), Eng. Geol. , 97( 3–4), 126– 145. https://doi.org/10.1016/j.enggeo.2007.12.009 Google Scholar CrossRef Search ADS   Bozzano F., Martino S., Naso G., Prestininzi A., Romeo R.W., Scarascia Mugnozza G., 2004. The large Salcito landslide triggered by the 2002 Molise, Italy, Earthquake, Earthq. Spectra , 20( S1), S95– S105. https://doi.org/10.1193/1.1768539 Google Scholar CrossRef Search ADS   Bozzano F., Cardarelli E., Cercato M., Lenti L., Martino S., Paciello A., Scarascia Mugnozza G., 2008. Engineering-geology model of the seismically-induced Cerda landslide (Sicily, Italy), Boll. Geofis. Teor. Appl. , 49( 2), 205– 225. Bozzano F., Lenti L., Martino S., Paciello A., Scarascia Mugnozza G., 2011. Evidences of landslide earthquake triggering due to self-excitation process, Int. J. Earth Sci. , 100( 4), 861– 879. https://doi.org/10.1007/s00531-010-0514-5 Google Scholar CrossRef Search ADS   Burjánek J., Gassner-Stamm G., Poggi V., Moore J.R., Fäh D., 2010. Ambient vibration analysis of an unstable mountain slope, Geophys. J. Int. , 180( 2), 820– 828. https://doi.org/10.1111/j.1365-246X.2009.04451.x Google Scholar CrossRef Search ADS   Burjánek J., Moore J.R., Yugsi Molina F.X., Fäh D., 2012. Instrumental evidence of normal mode rock slope vibration, Geophys. J. Int. , 188( 2), 559– 569. https://doi.org/10.1111/j.1365-246X.2011.05272.x Google Scholar CrossRef Search ADS   Chambers J.E. et al.  , 2011. Three-dimensional geophysical anatomy of an active landslide in Lias Group mudrocks, Cleveland Basin, UK, Geomorphology , 125( 4), 472– 484. https://doi.org/10.1016/j.geomorph.2010.09.017 Google Scholar CrossRef Search ADS   Clinton J.F., Case Bradford S., Heaton T.H., Favela J., 2006. The observed wander of the natural frequencies in a structure, Bull. seism. Soc. Am. , 96( 1), 237– 257. https://doi.org/10.1785/0120050052 Google Scholar CrossRef Search ADS   Colombero C., Baillet L., Comina C., Jongmans D., Vinciguerra S., 2017. Characterization of the 3-D fracture setting of an unstable rock mass: From surface and seismic investigations to numerical modeling, J. geophys. Res. , 122, 6346– 6366. https://doi.org/10.1002/2017JB014111 Google Scholar CrossRef Search ADS   Deere D.U., Miller R.P., 1966. Engineering Classification and Index Properties for Intact Rock , Illinois Univ. At Urbana Dept. of Civil Engineering. Del Gaudio V., Wasowski J., Pierri P., Mascia U., Calcagnile G., 2000. Gravimetric study of a retrogressive landslide in southern Italy, Surv. Geophys. , 21( 4), 391– 406. https://doi.org/10.1023/A:1006793009054 Google Scholar CrossRef Search ADS   Del Gaudio V., 2017. Instantaneous polarization analysis of ambient noise recordings in site response investigations, Geophys. J. Int. , 210( 1), 443– 464. https://doi.org/10.1093/gji/ggx175 Google Scholar CrossRef Search ADS   Del Gaudio V., Coccia S., Wasowski J., Gallipoli M.R., Mucciarelli M., 2008. Detection of directivity in seismic site response from microtremor spectral analysis, Nat. Hazards Earth Syst. Sci. , 8, 751– 762. https://doi.org/10.5194/nhess-8-751-2008 Google Scholar CrossRef Search ADS   Della Seta M., Martino S., Scarascia Mugnozza G., 2013. Quaternary sea-level change and slope instability in coastal areas: Insights from the Vasto Landslide (Adriatic coast, central Italy), Geomorphology , 201, 462– 478. https://doi.org/10.1016/j.geomorph.2013.07.019 Google Scholar CrossRef Search ADS   Dewez T.J.B., Rohmer J., Regard V., Cnudde C., 2013. Probabilistic coastal cliff collapse hazard from repeated terrestrial laser surveys: Case study from Mesnil Val (Normandy, northern France), J. Coast. Res. , Spec. Issue, 65, 702– 707. https://doi.org/10.2112/SI65-119.1 Google Scholar CrossRef Search ADS   Dickson M.E., Ogawa H., Kench P.S., Hutchinson A., 2013. Sea-cliff retreat and shore platform widening: steady-state equilibrium?, Earth Surf. Process. Landf. , 38( 9), 1046– 1048. https://doi.org/10.1002/esp.3422 Google Scholar CrossRef Search ADS   Doebling S.W., Farrar C.R., Prime M.B., Shevitz D.W., 1996. Damage Identification and Health Monitoring of Structural and Mechanical Systems from Changes in Their Vibration: A Literature Review. Technical Report LA-13070-MS , Los Alamos National Laboratory. Douze E.J., 1964. Rayleigh waves in short-period seismic noise, Bull. seism. Soc. Am. , 54( 4), 1197– 1212. https://doi.org/https://pubs.geoscienceworld.org/ssa/bssa/article/57/1/55/116458/short-period-seismic-noise Douze E.J., 1967. Short-period seismic noise, Bull. seism. Soc. Am. , 57( 1), 55– 81. Epifanio B., Zezere J.L., Neves M., 2013. Identification of hazardous zones combining cliff retreat rates with landslide susceptibility assessment, J. Coast. Res. , 65, 1681– 1686. https://doi.org/10.2112/SI65-284.1 Google Scholar CrossRef Search ADS   Farrugia D., Paolucci E., D’Amico S., Galea P., 2016. Inversion of surface wave data for subsurface shear wave velocity profiles characterized by a thick buried low-velocity layer, Geophys. J. Int. , 206( 2), 1221– 1231. https://doi.org/10.1093/gji/ggw204 Google Scholar CrossRef Search ADS   Farrugia D., Galea P., D’Amico S., Paolucci E., 2017. Sensitivity of ground motion parameters to local shear-wave velocity models: The case of buried low-velocity layers. Soil Dyn. Earthq. Eng. , 100, 196– 205. https://doi.org/10.1016/j.soildyn.2017.05.033 Google Scholar CrossRef Search ADS   Fiorucci M., Iannucci R., Lenti L., Martino S., Paciello A., Prestininzi A., Rivellino S., 2017. Nanoseismic monitoring of gravity-induced slope instabilities for the risk management of an aqueduct infrastructure in Central Apennines (Italy), Nat. Hazards , 86( S2), 345– 362. https://doi.org/10.1007/s11069-016-2516-5 Google Scholar CrossRef Search ADS   Galea P., 2007. Seismic history of the Maltese islands and considerations on seismic risk, Ann. Geophys. , 50( 6), 725– 740. https://doi.org/10.4401/ag-3053 Galea P., D’Amico S., Farrugia D., 2014. Dynamic characteristics of an active coastal spreading area using ambient noise measurements-Anchor Bay, Malta, Geophys. J. Int. , 199( 2), 1166– 1175. https://doi.org/10.1093/gji/ggu318 Google Scholar CrossRef Search ADS   Gigli G., Frodella W., Mugnai F., Tapete D., Cigna F., Fanti R., Intrieri E., Lombardi L., 2012. Instability mechanisms affecting cultural heritage sites in the Maltese Archipelago, Nat. Hazards Earth Syst. Sci. , 12( 6), 1883– 1903. https://doi.org/10.5194/nhess-12-1883-2012 Google Scholar CrossRef Search ADS   Glade T., Stark P., Dikau R., 2005. Determination of potential landslide shear plane depth using seismic refraction-a case study in Rheinhessen, Germany, Bull. Eng. Geol. Environ. , 64( 2), 151– 158. https://doi.org/10.1007/s10064-004-0258-1 Google Scholar CrossRef Search ADS   Got J.-L., Mourot P., Grangeon J., 2010. Pre-failure behaviour of an unstable limestone cliff from displacement and seismic data, Nat. Hazards Earth Syst. Sci. , 10( 4), 819– 829. https://doi.org/10.5194/nhess-10-819-2010 Google Scholar CrossRef Search ADS   Goudie A., 2004. Encyclopedia of Geomorphology , Routledge. Guzzetti F., Carrara A., Cardinali M., Reichenbach P., 1999. Landslide hazard evaluation: a review of current techniques and their application in a multi-scale study, Central Italy, Geomorphology , 31( 1-4), 181– 216. https://doi.org/10.1016/S0169-555X(99)00078-1 Google Scholar CrossRef Search ADS   Hack R., 2000. Geophysics for slope stability, Surv. Geophys. , 21( 4), 423– 448. https://doi.org/10.1023/A:1006797126800 Google Scholar CrossRef Search ADS   Hampton M.A., Griggs G.B., Edil T.B., Guy D.E., Kelley J.T., Komar P.D., Mickelson D.M., Shipman H.M., 2004. Processes that Govern the Formation and Evolution of Coastal Cliffs, in Formation, Evolution, and Stability of Coastal Cliffs-Status and Trends , pp. 1– 4, eds Hampton M.A., Griggs G.B., USGS - Professional Paper 1693. Havenith H.-B., Jongmans D., Abdrakmatov K., Trefois P., Delvaux D., Torgoev A., 2000. Geophysical investigations on seismically induced surface effects, case study of a landslide in the Suusamyr valley, Kyrgyzstan, Surv. Geophys. , 21, 349– 369. Havenith H.-B., Jongmans D., Faccioli E., Abdrakhmatov K., Bard P.-Y., 2002. Site effect analysis around the seismically induced Ananevo rockslide, Kyrgyzstan, Bull. seism. Soc. Am. , 92( 8), 3190– 3209. https://doi.org/10.1785/0120010206 Google Scholar CrossRef Search ADS   Havenith H.-B., Strom A., Jongmans D., Abdrakhmatov A., Delvaux D., Tréfois P., 2003a. Seismic triggering of landslides, Part A: Field evidence from the Northern Tien Shan, Nat. Hazards Earth Syst. Sci. , 3( 1/2), 135– 149. https://doi.org/10.5194/nhess-3-135-2003 Google Scholar CrossRef Search ADS   Havenith H.-B., Vanini M., Jongmans D., Faccioli E., 2003. Initiation of earthquake-induced slope failure: influence of topographical and other site specific amplification effects, J. Seismol. , 7( 3), 397– 412. https://doi.org/10.1023/A:1024534105559 Google Scholar CrossRef Search ADS   Hibert C., Grandjean G., Bitri A., Travelletti J., Malet J.-P., 2012. Characterizing landslides through geophysical data fusion: Example of the La Valette landslide (France), Eng. Geol. , 128, 23– 29. https://doi.org/10.1016/j.enggeo.2011.05.001 Google Scholar CrossRef Search ADS   Hoek E., Bray J.W., 1981. Rock Slope Engineering , Institution of Mining and Metallurgy. Horike M., 1985. Inversion of phase velocity of long-period microtremors to the S-wave-velocity structure down to the basement in urbanized areas, J. Phys. Earth , 33( 2), 59– 96. https://doi.org/10.4294/jpe1952.33.59 Google Scholar CrossRef Search ADS   Hutchinson J.N., 1988. General Report: Morphological and geotechnical parameters of landslides in relation to geology and hydrogeology, in Proceedings of Fifth International Symposium on Landslides , vol. 1, pp. 3– 35, ed. Bonnard C., Balkema. Google Scholar CrossRef Search ADS   Hutchinson J.N., Bromhead E.N., Chandler M.P., 1991. Investigations of landslides at St. Catherine's Point, Isle of Wight, in Slope Stability Engineering: Developments and Applications , pp. 169– 178, ed. Chandler M.P., Thomas Telford Publishing. Google Scholar CrossRef Search ADS   Hyde H.P.T., 1955. Geology of the Maltese Islands , Lux Press. Iannucci R., Martino S., Paciello A., D’Amico S., 2017. Rock mass characterization coupled with seismic noise measurements to analyze the unstable cliff slope of the Selmun Promontory (Malta), Procedia Eng. , 191, 263– 269. https://doi.org/10.1016/j.proeng.2017.05.180 Google Scholar CrossRef Search ADS   Imposa S., Grassi S., Fazio F., Rannisi G., Cino P., 2017. Geophysical surveys to study a landslide body (north-eastern Sicily), Nat. Hazards , 86( S2), 327– 343. https://doi.org/10.1007/s11069-016-2544-1 Google Scholar CrossRef Search ADS   ISRM, 1978. Suggested methods for the quantitative description of discontinuities in rock masses, Int. J. Rock Mech. Min. Sci. Geomech. Abstr. , 15, 319– 368. https://doi.org/10.1016/0148-9062(78)91472-9 CrossRef Search ADS   ISRM, 1979. Suggested method for determining water content, porosity, density, absorption and related properties and swelling and slake durability index properties, Int. J. Rock Mech. Min. Sci. Geomech. Abstr. , 16, 141– 156. Jongmans D., Garambois S., 2007. Geophysical investigation of landslides: A review, Bull. Soc. Geol. France , 178( 2), 101– 112. https://doi.org/10.2113/gssgfbull.178.2.101 Google Scholar CrossRef Search ADS   Kim J.T., Stubbs N., 2003. Crack detection in beam-type structures using frequency data, J. Sound Vib. , 259( 1), 145– 160. https://doi.org/10.1006/jsvi.2002.5132 Google Scholar CrossRef Search ADS   Konno K., Ohmachi T., 1998. Ground-motion characteristics estimated from spectral ratio between horizontal and vertical components of microtremor, Bull. seism. Soc. Am. , 88, 228– 241. Lenti L., Martino S., Paciello A., Prestininzi A., Rivellino S., 2012. Microseismicity within a karstified rock mass due to cracks and collapses as a tool for risk management, Nat. Hazards , 64( 1), 359– 379. https://doi.org/10.1007/s11069-012-0245-y Google Scholar CrossRef Search ADS   Lenti L., Martino S., Paciello A., Prestininzi A., Rivellino S., 2015. Recorded displacements in a landslide slope due to regional and teleseismic earthquakes, Geophys. J. Int. , 201( 3), 1335– 1345. https://doi.org/10.1093/gji/ggv063 Google Scholar CrossRef Search ADS   Lévy C., Baillet L., Jongmans D., Mourot P., Hantz D., 2010. Dynamic response of the Chamousset rock column (Western Alps, France), J. geophys. Res. , 115( F4), F04043, doi:10.1029/2009JF001606. https://doi.org/10.1029/2009JF001606 Google Scholar CrossRef Search ADS   Lévy C., Jongmans D., Baillet L., 2011. Analysis of seismic signals recorded on a prone-to-fall rock column (Vercors massif, French Alps), Geophys. J. Int. , 186( 1), 296– 310. https://doi.org/10.1111/j.1365-246X.2011.05046.x Google Scholar CrossRef Search ADS   Li T.M.C., Ferguson J.F., Herrin E., Durham H.B., 1984. High-frequency seismic noise at Lajitas, Texas, Bull. seism. Soc. Am. , 74( 5), 2015– 2033. Mantovani M., Devoto S., Forte E., Mocnik A., Pasuto A., Piacentini D., Soldati M., 2013. A multidisciplinary approach for rock spreading and block sliding investigation in the north-western coast of Malta, Landslides , 10( 5), 611– 622. https://doi.org/10.1007/s10346-012-0347-3 Google Scholar CrossRef Search ADS   Martino S., Mazzanti P., 2014. Integrating geomechanical surveys and remote sensing for sea cliff slope stability analysis: the Mt. Pucci case study (Italy), Nat. Hazards Earth Syst. Sci. , 14( 4), 831– 848. https://doi.org/10.5194/nhess-14-831-2014 Google Scholar CrossRef Search ADS   Martino S. et al.  , 2017. Ground effects triggered by the 24th August 2016, Mw 6.0 Amatrice (Italy) earthquake: surveys and inventorying to update the CEDIT catalogue, Geogr. Fis. Din. Quat. , 40, 77– 95. Maurer H., Spillmann T., Heincke B., Hauck C., Loew S., Springman S.M., Green A.G., 2010. Geophysical characterization of slope instabilities, First Break , 28( 8), 53– 61. McCann D.M., Forster A., 1990. Reconnaissance geophysical methods in landslide investigations, Eng. Geol. , 29( 1), 59– 78. https://doi.org/10.1016/0013-7952(90)90082-C Google Scholar CrossRef Search ADS   Méric O., Garambois S., Jongmans D., Wathelet M., Chatelain J.L., Vengeon J.M., 2005. Application of geophysical methods for the investigation of the large gravitational mass movement of Séchilienne, France, Can. Geotech. J. , 42( 4), 1105– 1115. https://doi.org/10.1139/t05-034 Google Scholar CrossRef Search ADS   Montoya-Montes I., Rodríguez-Santalla I., Sánchez-García M.J., Alcántara-Carrió J., Martín-Velázquez S., Gómez-Ortiz D., Martín-Crespo T., 2012. Mapping of landslide susceptibility of coastal cliffs: The Mont-Roig del Camp case study, Geol. Acta , 10, 439– 455. Moore J.R., Gischig V., Burjánek J., Loew S., Fäh D., 2011. Site effects in unstable rock slopes: dynamic behavior of the Randa instability (Switzerland), Bull. seism. Soc. Am. , 101( 6), 3110– 3116. https://doi.org/10.1785/0120110127 Google Scholar CrossRef Search ADS   Moore J.R. et al.  , 2016. Anthropogenic sources stimulate resonance of a natural rock bridge, Geophys. Res. Lett. , 43( 18), 9669– 9676. https://doi.org/10.1002/2016GL070088 Google Scholar CrossRef Search ADS   Nakamura Y., 1989. A method for dynamic characteristics estimation of subsurface using microtremor on the ground surface, Q. Rep. Railw. Tech. Res. Inst. , 30( 1), 25– 33. Nogoshi M., Igarashi T., 1970. On the propagation characteristics of microtremors, J. Seismol. Soc. Japan , 23, 264– 280. (in Japanese with English abstract). Nogoshi M., Igarashi T., 1971. On the amplitude characteristics of microtremors, J. Seismol. Soc. Japan , 24, 24– 40. (in Japanese with English abstract). Panzera F., D’Amico S., Lotteri A., Galea P., Lombardo G., 2012. Seismic site response of unstable steep slope using noise measurements: the case study of Xemxija Bay area, Malta, Nat. Hazards Earth Syst. Sci. , 12( 11), 3421– 3431. https://doi.org/10.5194/nhess-12-3421-2012 Google Scholar CrossRef Search ADS   Panzera F., D’Amico S., Galea P., Lombardo G., Gallipoli M.R., Pace S., 2013. Geophysical measurements for site response investigation: preliminary results on the island of Malta, Boll. Geofis. Teor. Appl. , 54, 111– 128. Panzera F., Pischiutta M., Lombardo G., Monaco C., Rovelli A., 2014. Wavefield polarization in fault zones of the Western flank of Mt. Etna: Observations and fracture orientation modelling, Pure appl. Geophys. , 171( 11), 3083– 3097. https://doi.org/10.1007/s00024-014-0831-x Google Scholar CrossRef Search ADS   Pazzi V., Tanteri L., Bicocchi G, D’Ambrosio M., Caselli A., Fanti R., 2017. H/V measurements as an effective tool for the reliable detection of landslide slip surfaces: Case studies of Castagnola (La Spezia, Italy) and Roccalbegna (Grosseto, Italy), Phys. Chem. Earth, Parts A/B/C , 98, 136– 153. https://doi.org/10.1016/j.pce.2016.10.014 Google Scholar CrossRef Search ADS   Pedley M., 2011. The Calabrian Stage, Pleistocene highstand in Malta: a new marker for unravelling the Late Neogene and Quaternary history of the islands, J. Geol. Soc. , 168( 4), 913– 926. https://doi.org/10.1144/0016-76492010-080 Google Scholar CrossRef Search ADS   Pedley H.M., House M.R., Waugh B., 1976. The geology of Malta and Gozo, Proc. Geologists' Assoc. , 87( 3), 325– 341. https://doi.org/10.1016/S0016-7878(76)80005-3 Google Scholar CrossRef Search ADS   Pedley H.M., House M.R., Waugh B., 1978. The geology of the Pelagian block: the Maltese islands, in The Ocean Basins and Margins , pp. 417– 433, eds Nairn A.E.M., Kanes W.H., Stehli F.G., Plenum Press. Google Scholar CrossRef Search ADS   Pedley M., Clarke M., Galea P., 2002. Limestone Isles in a Crystal Sea: The Geology of the Maltese Islands , Publishers Enterprises Group Ltd. Pennetta M., Russo E.L., 2011. Hazard factors in high rocky coasts of Capri Island (Gulf of Naples, Italy), J. Coast. Res. , 61, 428– 434. https://doi.org/10.2112/SI61-001.53 Google Scholar CrossRef Search ADS   Petley D., 2012. Global patterns of loss of life from landslides, Geology , 40( 10), 927– 930. https://doi.org/10.1130/G33217.1 Google Scholar CrossRef Search ADS   Pischiutta M. et al.  , 2017. Results from shallow geophysical investigations in the northwestern sector of the island of Malta, Phys. Chem. Earth, Parts A/B/C , 98, 41– 48. https://doi.org/10.1016/j.pce.2016.10.013 Google Scholar CrossRef Search ADS   Ramamurthy T., 2004. A geo-engineering classification for rocks and rock masses, Int. J. Rock Mech. Min. Sci. , 41( 1), 89– 101. https://doi.org/10.1016/S1365-1609(03)00078-9 Google Scholar CrossRef Search ADS   Rosser N.J., Brain M.J., Petley D.N., Lim M., Norman E.C., 2013. Coastline retreat via progressive failure of rocky coastal cliffs, Geology , 41( 8), 939– 942. https://doi.org/10.1130/G34371.1 Google Scholar CrossRef Search ADS   Sass O., Bell R., Glade T., 2008. Comparison of GPR, 2D-resistivity and traditional techniques for the subsurface exploration of the Öschingen landslide, Swabian Alb (Germany), Geomorphology , 93( 1-2), 89– 103. https://doi.org/10.1016/j.geomorph.2006.12.019 Google Scholar CrossRef Search ADS   Senfaute G., Duperret A., Lawrence J.A., 2009. Micro-seismic precursory cracks prior to rock-fall on coastal chalk cliffs: a case study at Mesnil-Val, Normandie, NW France, Nat. Hazards Earth Syst. Sci. , 9( 5), 1625– 1641. https://doi.org/10.5194/nhess-9-1625-2009 Google Scholar CrossRef Search ADS   Sitharam T.G., Sridevi J., Shimizu N., 2001. Practical equivalent continuum characterization of jointed rock masses, Int. J. Rock Mech. Min. Sci. , 38( 3), 437– 448. https://doi.org/10.1016/S1365-1609(01)00010-7 Google Scholar CrossRef Search ADS   Starr A.M., Moore J.R., Thorne M.S., 2015. Ambient resonance of Mesa Arch, Canyonlands National Park, Utah, Geophys. Res. Lett. , 42( 16), 6696– 6702. https://doi.org/10.1002/2015GL064917 Google Scholar CrossRef Search ADS   Tang C., Li L., Xu N., Ma K., 2015. Microseismic monitoring and numerical simulation on the stability of high-steep rock slopes in hydropower engineering, J. Rock Mech. Geotech. Eng. , 7( 5), 493– 508. https://doi.org/10.1016/j.jrmge.2015.06.010 Google Scholar CrossRef Search ADS   Toksöz M.N., Lacoss R.T., 1968. Microseisms: mode structure and sources, Science , 159( 3817), 872– 873. https://doi.org/10.1126/science.159.3817.872 Google Scholar CrossRef Search ADS PubMed  Valentin J., Capron A., Jongmans D., Baillet L., Bottelin P., Donze F., Larose E., Mangeney A., 2017. The dynamic response of prone-to-fall columns to ambient vibrations: comparison between measurements and numerical modelling, Geophys. J. Int. , 208( 2), 1058– 1076. https://doi.org/10.1093/gji/ggw440 Google Scholar CrossRef Search ADS   Varnes D.J., 1978. Slope movement types and processes, in Landslides Analysis and Control, Special Report 176 , pp. 11– 33, eds Schuster R.L., Krizek R.J., National Academy of Science. Varnes D.J. & IAEG, 1984. Landslide Hazard Zonation: A Review of Principles and Practice , United Nations Scientific and Cultural Organization. Vella A., Galea P., D’Amico S., 2013. Site frequency response characterisation of the Maltese islands based on ambient noise H/V ratios, Eng. Geol. , 163, 89– 100. https://doi.org/10.1016/j.enggeo.2013.06.006 Google Scholar CrossRef Search ADS   Vidale J.E., 1986. Complex polarisation analysis of particle motion, Bull. seism. Soc. Am. , 76, 1393– 1405. Xu N.W., Li T.B., Dai F., Li B., Zhu Y.G., Yang D.S., 2015. Microseismic monitoring and stability evaluation for the large scale underground caverns at the Houziyan hydropower station in Southwest China, Eng. Geol. , 188, 48– 67. https://doi.org/10.1016/j.enggeo.2015.01.020 Google Scholar CrossRef Search ADS   Yalcinkaya E. et al.  , 2016. Near-surface geophysical methods for investigating the Buyukcekmece landslide in Istanbul, Turkey, J. Appl. Geophys. , 134, 23– 35. https://doi.org/10.1016/j.jappgeo.2016.08.012 Google Scholar CrossRef Search ADS   Yamanaka H., Takemura M., Ishida H., Niwa M., 1994. Characteristics of long-period microtremors and their applicability in exploration of deep sedimentary layers, Bull. seism. Soc. Am. , 84( 6), 1831– 1841. Young A.P., Ashford S.A., 2008. Instability investigation of cantilevered seacliffs, Earth Surf. Process. Landf. , 33( 11), 1661– 1677. https://doi.org/10.1002/esp.1636 Google Scholar CrossRef Search ADS   Zang A. et al.  , 2017. Hydraulic fracture monitoring in hard rock at 410 m depth with an advanced fluid-injection protocol and extensive sensor array, Geophys. J. Int. , 208( 2), 790– 813. https://doi.org/10.1093/gji/ggw430 Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of The Royal Astronomical Society.

### Journal

Geophysical Journal InternationalOxford University Press

Published: May 1, 2018

## You’re reading a free preview. Subscribe to read the entire article.

### DeepDyve is your personal research library

It’s your single place to instantly
that matters to you.

over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month ### Explore the DeepDyve Library ### Unlimited reading Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere. ### Stay up to date Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates. ### Organize your research It’s easy to organize your research with our built-in tools. ### Your journals are on DeepDyve Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more. All the latest content is available, no embargo periods. ### Monthly Plan • Read unlimited articles • Personalized recommendations • No expiration • Print 20 pages per month • 20% off on PDF purchases • Organize your research • Get updates on your journals and topic searches$49/month

14-day Free Trial

Best Deal — 39% off

### Annual Plan

• All the features of the Professional Plan, but for 39% off!
• Billed annually
• No expiration
• For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588$360/year

billed annually

14-day Free Trial