Studies of long-term seismic activity in any region of the world require the longest historical perspective as possible. In most of the regions of Europe, before the XIIIth and the XIVth centuries, historical sources allow to establish a catalogue of felt earthquakes but the available data give very few information to estimate local intensities and epicentres.

Thus, the knowledge of the seismic activity begins really at this epoch and covers a time period of less than 700 years. From that time, sources from different origins. (chronicles, annotations, parish registers, account registers,…) give us more details on local effects and allow reliable estimation of damages and perceptibility area of the earthquakes. It is then possible to determine the probable epicentral area of the strongest earthquakes and also to estimate their magnitude by comparison with recent earthquakes for which the magnitude has been instrumentally determined.

The Bree Uplift is a particular structure in the direct footwall of the southwestern graben boundary fault system of the Roer Valley Graben, which has been firstly recognized at the base of the Cretaceous. To date fault activity around the Bree Uplift has been confined to the Subhercynian (late Cretaceous) inversion event or considered fading out during Tertiary times.

The revision of the existing geological data reveals that the Bree Uplift can still be recognized on the top-Tertiary map. This infers at least a late Tertiary activity, suggesting continuous fault activity in the graben boundary fault system not only on the major boundary faults but also on different splay faults, bounding individual fault blocks.

A high-resolution reflection seismic survey was carried out in 1999 over the Feldbiss fault system, the southern border of the Roer Valley graben, in Belgium. Six profile-lines with total length of 13982 m provided information on the 40-600 m depth range, covering Lower Pleistocene to Miocene strata with special emphasis on the Plio-Pleistocene Kieseloolite formation. Data quality depends on near-surface conditions and on degree of deformation in some fault zones, with better results for seismic detonator sources compared to vibroseis sources. The new data confirm the segmented character of the fault system with occurrence of fault bends, relay ramps and branching of overlapping fault sequences, testifying of the strong tectonic activity during the lower Pleistocene. Antiform structures along the Bichterweerd scarp, relaying the Feldbiss to the Geleen fault in the Meuse valley, are presented as a model for the Tertiary evolution of the Bree Uplift.

We studied the applicability of classical scarp degradation modelling to active normal faults in the Lower Rhine Embayment. Our quantitative analysis was conducted on the frontal Bree fault scarp (Feldbiss fault) in Belgium and the Peel fault scarp near the city of Neer in the Netherlands. Vertical offset and diffusion age of these scarps have been modelled from elevation profiles across the studied faults using the diffusion equation. For that purpose, a computer-program (profil 2000) has been written, providing a sensitivity analysis of the determined parameters in function of the spatial repartition of the elevation measurements along the considered profiles. The results of this morphometric analysis have been validated by a comparison with the geologic record of the tectonic activity observed in the trenches excavated at the sites where the measurements have been conducted.

We conclude that the modelling can only be applied to study tectonic activity since the Last Glacial Maximum (±14-19 kyr BP) because the surface expression of older paleoearthquakes in unconsolidated Late Pleistocene sediments has been erased by the strong erosive phase that occurred at the end of this glacial period. Even for Holocene scarps, morphologic dating seems very difficult because man-made perturbations destroyed surface evidence of the very recent fault activity in many sites. Nevertheless, we estimate that an appropriate value for the mass diffusivity constant for~ 1-m-high scarps in the investigated region is 0.002 to 0.010 m2/yr. On the other hand, vertical offsets can be determined with a good precision. These amount to respectively ~1 m and 1,3 m since the Last Glacial Maximum on the Feldbiss fault in Belgium and the Peel fault near Roermond in the Netherlands.

Topographic and morphologic models based on detailed Digital Elevation Models (DEM) of the Rur Graben, in particular a 33 km section of the Rurrand Fault, proved to be essential in preliminary investigations for the site selection of a trench for detailed paleoseismological investigations. The entire DEM in the 33 km Jülich-Düren area displayed as a color-shaded terrain map clearly illustrates the main features of the eastern border fault of the Rur Graben. As investigations concentrated on the 8x6 km Stetternich-Hambach section of the Rurrand fault, the use of contoured topographic, gray-shaded terrain and terrain slope maps helped delineate the main features of the surface fault expression. On the basis of topographic profiles constructed from the DEM, no fault scarp could be identified; however, the detailed surveying results from two geophysical profiles correlated well with the topographic data.

In preparation of the first paleoseismic trenching in the NE border of the Roer graben (the Netherlands), site selection was carried out. Combining geological and seismological information and using existing aerial photographs, seismic reflection and geodetic levelling data, it was decided to focus on the Peel boundary fault near the village of Neer. Detailed information on the exact location of the fault was obtained through geophysical techniques, mainly ground penetrating radar (GPR) and resistivity measurements. GPR data unambiguously showed the flexuring and offset of reflectors affected by the fault. Performing eleven GPR profiles along strike allowed to obtain a 3D picture of the fault, laterally extending the information given in the trench.

The Carboneras Fault Zone (CFZ) represents an active set of sinistral strike-slip faults in the Betic Cordilleras of southeastern Spain. It constitutes a major segment of the ‘Trans-Alboran shear zone’ during the Cenozoic, striking NE-SW. The CFZ separates the Cabo de Gata Block (Neogene volcanics) against Neogene basinal sediments and the metamorphic basement of the Alpujarride Complex.

Three sites along the CFZ were examined with Ground Penetrating Radar techniques. Radar surveying was complemented by structural studies. Shallow-depth high-resolution imaging of Tyrrhenian beach terraces exhibited both vertical and minor horizontal offsets in the Rambla Morales site in the south. A sinistral strike-slip fault associated with minor thrust faults in a positive flower structure was detected in the middle segment along the La Serrata ridge, sealed by a caliche of late Pleistocene age (> 10 ka). The Playa de Bolmayor section yielded sub-surface evidence for several faults probably related to recent activity of individual fault strands. Our results suggest a distributed tectonic activity of the CFZ during the Late Quaternary.

From 1998 to 2000, we have studied the evidence for large paleoearthquakes at the Rurrand Fault. This fault represents the eastern border of the Roer Valley Graben, which is the tectonically most active region in the Lower Rhine Embayment. The purpose of our paleoseismological studies is to enlarge the seismicity data base for this region beyond instrumental records and historical reports using indications of surface-faulting events from stratigraphie conditions at active faults. Larger time spans considered in the earthquake catalogue will enable a more reliable statistical analysis which is required for seismic hazard assessment. Based on analyses of geological data and géomorphologie investigations, detailed geophysical surveying was carried out along the southern Rurrand Fault segment for the selection of a site appropriate to paleoseismological studies. Mapping of physical parameter contrasts with seismic reflection, VES, ERT, and GPR measurements along fault-crossing profiles inferred position and near-surface structure of the fault. At the site promising the best conditions, a trench was excavated across the fault near the city of Jiilich, Germany. Within a depth of about 4 m, the Rurrand Fault was exposed in an about 50 m-wide system of faults and fault zones, affecting the stratigraphie sequence with various displacement characteristics and amounts of throw. According to heavy mineral analyses, the deposition time of most the exposed sediment strata was assigned to Pliocene and Lower Pleistocene time. These geological units are covered by loess layers deposited through so-lifluction processes during the Weichselian glacial, i.e. some tens of ka B.P., or – with lower probability – during the Saalian glacial. Several faults which had also affected the loess reflect younger fault activity. However, clear paleoseismic features were not observed in the trench, thus an unambiguous proof of the occurrence of coseismic fault displacements could not be furnished. Recently, differential subsidence due to drainage takes place in the surroundings of the nearby opencast mining. An amount of some 0.35 m, concentrated in a very narrow lateral zone, has been observed during the last 40 a at about 1 km distance from the trench position. To date, the subsidence could not be clearly located in the trench exposure. Results from geodetic levelling campaigns will help to determine the offset residuals and to gain better insight into the ruling displacement processes at the Rurrand Fault.

A first trench has been excavated for paleoseismological analysis in the German part of the Roer Valley graben, which has experienced several historical earthquakes with a maximum intensity up to VIII on the MSK-scale.The trench has exposed the Rurrand fault as a complex fault zone with at least five separate, SW-dipping, normal fault strands displacing an early Pleistocene terrace of the Rhine river by more than 7 m. The major part of the observed deformation was produced during or after deposition of an overlying unit of stratified loess of middle Weichselian to probably Saalian age. The faulting history is shown to be episodic, with different fault strands active at different times. Growth faulting that would be indicative of continuous, aseismic fault motion has not been observed. Our stratigraphic control is not sufficient to constrain the timing and to provide evidence of the coseismic nature for each observed fault displacement. However, two units of structureless, gravelly loess are interpreted as the result of extensive solifluction triggered by two large surface-rupturing events. This is suggested by the position of these units, which is controlled by the main faults, and by their remarkably young age (< 400 cal. BC), indicated by radiocarbon and OSL datings and by the presence of historic brick fragments. At least two faults show moderate activity that is even younger. Our interpretation is not in agreement with earlier hypotheses that ongoing vertical movements of circa 1 mm/a in the German part of the Lower Rhine graben are the result of aseismic fault creep, but is in line with the results of similar investigations on the southwestern border fault of the Roer Valley Graben in Belgium, which demonstrates the need for further paleoseismological research in this region. The Rurrand fault is presently experiencing aseismic slip on its superficial portion, induced by extensive groundwater lowering for mining purposes. This ongoing deformation seems to be expressed in the trench as diffuse bundles of anastomosing cracks extending up to, and in some cases even into the plough zone, rather than as sharp fault planes which are typical of older, tectonic fault movements.

The coversands along the Bree fault escarpment, NE Belgium, were investigated by a combined dating approach including infrared optically stimulated luminescence (IRSL), thermoluminescence (TL) and radiocarbon methods. Four trenches were excavated cutting the fault scarp near the village of Bree in northeast Belgium. Altogether 17 luminescence samples and seven radiocarbon samples were investigated in order to set up a more reliable and precise chronological frame for the local coversand stratigraphy and the timing of Late Quaternary earthquake events. The chronological results indicate at least five accumulation periods for the coversand units in the area of interest. The oldest coversands were deposited during Saalian or Early Weichselian followed by coversands deposited during the Early and/or Middle Weichselian. At least three coversand units can be distinguished for the time span of the Late Weichselian to Holocene. Significant TL age underestimation of more than 20% compared to IRSL was found for samples from Trench 4. Most of the IRSL age estimates are in agreement with radiocarbon dates from the same section.

Paleoseismological investigations have been performed at Mt. Baldo and in the Lessini Mts. in order to collect quantitative data on the activity of minor faults showing geomorphic evidence of recent activation. The 4.5-km-long, NNE-SSW trending Naole fault was responsible for the formation of a narrow depression at the top of Mt. Baldo, bordered by a continuous bedrock (carbonate) fault scarp to the west. The extensional activity along this minor fault is probably due to gravitational deformations (lateral spreading) in response to the warping of the Mt. Baldo anticline. A 1.5-km-long graben is instead related to the 2.5-km-long, NNW-SSE trending Orsara fault (Lessini Mts.) which was responsible for the formation of bedrock (carbonate) fault scarps. This minor fault is part of a complex structural framework made of few-km-long faults which show evidence of Quaternary activity. Two trenches have been excavated across the Naole fault which showed the occurrence of displacement events subsequent to 17435-16385 BP (cal. age) and probably prior to 5455-5385/5330-5295 BP (cal. age). Two other trenches have been excavated across the Orsara fault whose analysis indicated that the most recent displacement event occurred between 20630-19795 BP and 765-675 BP (cal. age). The upper chronological limits of the displacements give some indications about the minimum elapsed time since the last fault activation (about 5,300 years for the Naole fault and 5-8 centuries for the Orsara fault). Both 1) the maximum expected magnitude of the earthquakes which may originate along the Mt. Baldo thrust and 2) the identification of a main fault responsible for the displacements along the complex net of minor faults affecting the Lessini Mts. are still open questions. As for point 1 although historical earthquakes with magnitude 4.5-5 may be associated with the Mt. Baldo thrust, the investigations carried out in this area did not clarify whether larger magnitude earthquakes may be expected. As for point 2, the cause of the displacements along the Orsara (Lessini Mts.) fault may be related to the activity of a major blind fault (which, however, has never been identified), responsible for the uplift of the Lessini Mts. More generally, the obtained results demonstrate the limits of traditional paleoseismological analyses in Alpine areas whose erosional/depositional activity has been strongly conditioned by the Late Pleistocene glacial history. The lack of units younger than loess and colluvial sediments related to the Last Glacial Maximum makes it impossible to define narrower chronological constraints for the displacements and to estimate the number and size of the displacement events. Moreover, the rebound following the retreat of the thick glacial cover affecting the Alpine area may have induced stresses responsible for higher deformation rates after the Last Glacial Maximum. Higher surficial deformation rates could imply shorter recurrence intervals for faulting episodes and/or larger magnitude earthquakes. Therefore, paleoseismologically inferred data in Alpine areas may not correctly define the fault behaviour related to the present tectonic regime.

Paleoseismological techniques have been used to investigate gravitational deformations at the Mortirolo Pass (Valtellina region, central Alps), in order to improve the knowledge on the activation mechanisms and the evolution of deep-seated gravitational slope movements. The deformation has been responsible for mass sliding towards the Valtellina depression through the activation of several-hundred-metre-long shear planes. Minor shear planes dipping towards the mountain played the role of antithetic structures. Four trenches were excavated across scarps representing the surficial expression of shear planes affecting the bedrock and Late Pleistocene-Holocene deposits. The excavations enabled to investigate the stratigraphy of Quaternary deposits and the geometry and kinematics of the shear planes affecting them. Radiocarbon analyses on organic material contained in sediments and paleosols enabled to define a succession of displacement events which occurred during the Late Pleistocene-Holocene. Collected data indicate the persistence of the activity until recent times (last movement related to 1810-1540 cal. BP). A sudden movement has been detected along one of the main shear surfaces (dipping towards the valley) with a vertical displacement of several metres. In contrast, numerous displacements (with lower vertical offset) have been detected along the antithetic shear planes. Different hypotheses have been proposed in the past to define the origin of huge gravitational movements (glacial retreat, uplift of the Alpine chain, fault activity). However, the Late Pleistocene cycles of glacial loading and unloading on the mountain slopes seem to be the most probable factors causing deep-seated gravitational movements in the investigated region. A recent dramatic landslide in an area adjacent to the investigated one (Mt. Zandila-Valpola) testifies to the paroxistic evolution of the large scale gravitational deformations. The densely inhabited Valtellina region is affected by a large number of gravitational structures similar to those of the Mortirolo area. In consideration of the possible effects of the paroxistic activation of these structures, detailed studies on the chronology and kinematics of the deformations through the application of paleoseismological techniques should therefore be encouraged.

Given that earthquakes do not occur only along high slip-rate faults, slow moving seismogenic faults should be characterized in order to minimize seismic hazard uncertainties. Although no historical earthquakes related to the El Camp fault have been documented, earlier regional geological analysis and the presence of a fault scarp provide evidence of its activity. A paleoseismological study on the southern part of the fault was performed in accordance with the following steps: 1) geological and géomorphologie analysis focussing on the detection of evidence for the seismogenic behavior of the fault, 2) near fault analysis to select the best trenching sites, and 3) trenching to establish and characterize the paleoseismic events. Different dating procedures were used in regional and trenching analyses (Thermoluminiscence, U/Th, Radiocarbon, Pollen analysis). The seismogenic nature of the fault was established by the presence of liquefaction features related to the fault, and by the presence of colluvial wedges composed of large angular blocks. We identified a segment boundary to the north of the Porquerola creek and we focussed our attention on the southern segment, which was active after 125000 yr. The slip rate in this southern segment is 0.02 mm/yr. A minimum of three seismic events were detected, from young to old: the last event Z took place some time prior to 1195 yr AD, the penultimate event Y between event X and the Holocene, and, finally, event X occurred after 125000 yr and prior to 60000 yr. The recurrence period is between 25000 and 35000 yr, the elapsed time is estimated to be no longer than 3000 yr; and the maximum estimated earthquake considering both the onshore and the offshore part of the fault is Mw 6.7 +/-0,5.

The Amer fault is a 30 km long normal fault, which generated the damaging earthquakes of March and May 1427. Triangular facets, wine glass drainage basins, alluvial fans and scarps along the Amer fault mountain front provide evidence of its recent activity. Topographic profiling, electrical logging, tomographic and high-resolution seismic profiling along the northern segment of the Amer fault showed the following: i) no evidence of surface deformation in recent deposits; ii) fault scarps produced by the Amer fault located only on old alluvial fans, probably Pleistocene in age, and iii) Amer fault related deformation reaching upper Quaternary levels, but not the uppermost horizons. The high sedimentation rate (nearly one order of magnitude greater than the fault slip rate) due to the filling of the lake, which resulted from the damming of the Fluvià river by the Bosc de Tosca lava flow (17,000 yr BP), can account for the absence of surface deformation on Holocene sediments.

The Southern Alps – Ligurian basin junction is one of the most active seismic areas in Western Europe countries. The topographic and the structural setting of this region is complex because of (i) its position between the high topography of the Southern Alps and the deep, narrow Ligurian oceanic basin, and (ii) the large number of structures inherited from the Alpine orogeny. Historical seismicity reveals about twenty moderate-size earthquakes (up to M=6.0), mostly distributed along the Ligurian coast and the Vésubie valley. A recent geodetic experiment shows a significant strain rate during the last 50 years in the area between the Argentera massif and the Mediterranean coastline. Results of this experiment suggest a N-S shortening of about 2-4 mm/yr over the network, this shortening direction is consistent with the seismological (P-axes of earthquakes) and the microtectonic data. The Pennic front (E-NE of the Argentera massif) and the northern Ligurian margin are the most seismically active areas. In the Nice arc and in the Argentera massif, some seismic lineaments correspond to faults identified in the field (such as theTaggia-Saorge fault or the Monaco-Sospel fault). In the western part of the Alpes Maritimes, no seismic activity is recorded in the Castellane arc. In the field, geological evidence, such as offsets of recent alluvial sediments, recent fault breccia, speleothem deformations, radon anomalies and others indicates recent deformation along these faults. Nevertheless, to this date active fault scarps have not been identified: this probably results from a relatively high erosion rate versus deformation rate and from the lack of Quaternary markers. We also suspect the presence of two hidden active faults, one in the lower Var valley (Nice city area) and the other one at the base of the Argentera crustal thrust-sheet. Offshore, along the northern Ligurian margin, the seismic reflection data shows traces of Quaternary extensional deformation, but the accuracy of the data does not yet allow the construction of a structural map nor does it allow the determination of the continuity between the offshore and onshore structures. From these data set we propose a preliminary map of 11 active faults and we discuss the questions which remain unsolved in the perspective of seismic hazard evaluations.

An inventory of the available surficial data on active faults in Italy has been compiled by gathering all the available information on peninsular Italy (project by CNR, National Group for the Defense against Earthquakes), the central-eastern Alps and the Po Plain (EC ‘PALEOSIS’ project). Such information has been summarised in maps (reporting surficial expressions of faults with length L≥11 km) and in a table where fault parameters relevant for seismic hazard assessment (e.g. slip rates, recurrence intervals for surface faulting events, etc..) have been reported. Based on the geological characteristics of the Italian territory, a fault has been considered as active if it shows evidence of Late Pleistocene-Holocene displacements. Active faults in Italy are distributed throughout the entire Apennine chain, in the Sicilian and Calabrian regions and in some Alpine sectors, but knowledge is not homogeneously distributed through the territory. The largest amount of data is related to the central Apennines. In contrast, fault geometries and parameters are less well defined in the southern Apennines, Sicily and Calabria, where investigations have started more recently. Knowledge is sparse in the northern Apeninnes, where data necessary to define fault parameters are lacking and also the chronology of the activity has to be considered cautiously. Abundant blind faulting in the Po Plain hinders the detection of active faults by means of the classical surficial investigations and therefore the present knowledge is limited to the Mantova fault. Blind faults and the peculiar recent geological history of the Alpine areas, which is strongly conditioned by the erosional and depositional activity during and after the last glacial maximum, also hinder the identification of active faults in the central-eastern Alps. Some faults in this Alpine sector are believed to be active, but data on their segmentation are still missing. Available information indicates that Italian active faults are usually characterised by slip rates lower than 1 mm/yr. Recurrence intervals for surface faulting events are longer than 1,000 years in the central and southern Apennines. This review on the Italian active faults represents the first step to produce a map of the major seismic sources in Italy, which in turn will result from the merge of surficial data with seismological and geological subsurficial data. The available knowledge gathered in this paper indicates those areas where data are presently sparse. It should be, therefore, possible to better plan future geomorphological and paleoseismological investigations.

This paper presents observations of recent faulting activity in the karstic network of the Rochefort Cave (Namur Province, Belgium, Europe). The principal recent tectonic features are bedding planes reactivated as normal faults, neo-formatted normal faults in calcite flowstone, fresh scaling, extensional features, fallen blocks and displacement of karstic tube. The seismo-tectonic aspect is expanded by the presence of fallen blocks where normally the cavity must be very stable and in equilibrium. Three main N 070° fault planes and a minor one affect, at a decimetre scale, the karst features and morphology. The faults are still active because recent fresh scaling and fallen blocks are observable. The breaking of Holocene soda straw stalactites and displacements of artificial features observed since the beginning of the tourist activity, in the last century, also suggest very recent reactivation of these faults. This recent faulting can be correlated to present-day tectonic activity, already evidenced by earthquakes in the neighbouring area. Therefore, karstic caves are favourable sites for the observation and the quantification of recent tectonic activity because they constitute a 3-D framework, protected from erosion. Fault planes with this recent faulting present slickensides. Thus a quantitative analysis in term of stress inversion, with the help of striated faults, has permitted to reconstruct the stress tensor responsible for the brittle deformation. The principal NW-SE extension (σ3 horizontal) is nearly perpendicular to that of the present regional stress as illustrated by the analysis of the last strong regional earthquake (Roermond, The Netherlands) in 1992. During the Meso-Cenozoic, the main stress tectonics recorded in this part of the European platform is similar to the present one with a NE-SW direction of extension.

The discrepancy between the regional stress field and the local stress in the Rochefort cave can be the result of the inversion of the σ2 and σ3 axes of the stress ellipsoid due to its symmetry or of a local modification at the ground surface of the crustal stress field as it has been already observed in active zones.

Seismic hazard in low seismicity areas of Europe has traditionally been considered insignificant. However, in the light of the recently conducted paleoseismic studies along the Rhine Graben, a revision is required. Previously applied standard probabilistic seismic hazard assessment (PSHA) methods, using Poissonian approach for the earthquake occurrence, can now be substituted by renewal models where fault parameters such as the maximum magnitude, recurrence interval and the elapsed time since the last occurrence of a large earthquake, can be utilized. In this study, the application and the influence of the available paleoseismic data in the Lower Rhine Graben to seismic hazard analysis is demonstrated. The resulting hazard maps, when compared to the standard PSHA using Poissonian approach, indicate a more precise geographical distribution of the estimated seismic hazard levels. The influence of the paleoseismic data seem to be less important for return periods less than a 1000 years. Among the different input models, the highest values reach to 170 cm/sec2 for a 1000 year return period using a combination of Poissonian and renewal models.

To provide quantitative information on the ground acceleration necessary to break speleothems, laboratory measurements on samples of stalagmite have been performed to study their failure in bending. Due to their high natural frequencies, speleothems can be considered as rigid bodies to seismic strong ground motion. Using this simple hypothesis and the determined mechanical properties (a minimum value of 0.4 MPa for the tensile failure stress has been considered), modelling indicates that horizontal acceleration ranging from 0.3 m/s2 to 100 m/s2 (0.03 to 10g) are necessary to break 35 broken speleothems of the Hotton cave for which the geometrical parameters have been determined. Thus, at the present time, a strong discrepancy exists between the peak accelerations observed during earthquakes and most of the calculated values necessary to break speleothems. One of the future research efforts will be to understand the reasons of the defined behaviour. It appears fundamental to perform measurements on in situ speleothems.

In some karstic caves, the observation of numerous broken stalagmites may provide potential secondary evidence for intense palaeoearthquakes during recent past times. We have named these morpho-sedimentologic features seismothems. A methodology has been developed to discriminate broken speleothems due to earthquake-induced effects or caused by other mechanisms. A study has been carried out in the Belgian karst areas. In the Vesdre Valley, it seems difficult to find evidence of the well-known Verviers earthquake, probably the most destructive historical earthquake known in Belgium which occurred in 1692 AD. The most important concentration of broken stalagmites was discovered in the caves between Hotton and Han-sur-Lesse. The observations in the cave of Hotton suggest a seismic origin, the other origins can not be the cause of the speleothem break. This result implies a strong earthquake situated close to the cave. A preliminary AMS 14C age suggests a minimum age of 10100 ±1200 cal 14C yr BR for one stalagmite rupture in the Hotton cave.