3B2v8:06a=w ðDec 5 2003Þ:51c JQSR : 1459 Prod:Type:COM ED:Sushma XML:ver:5:0:1 pp:1223ðcol:fig::1;3;4;6;7;8;9;10Þ PAGN:NMN SCAN:Anil ARTICLE IN PRESS 1 3 Quaternary Science Reviews ] (]]]]) ]]]–]]] 5 7 First appraisal to define prospective seismogenic sources from 9 historical earthquake damages in southern Upper Rhine Graben 11 Umberto Fracassia,Ã, Bertrand Nivie´reb, Thierry Wintera 13 aBureau de Recherches Ge´ologiques et Minie´res, Div. ARN/MAS, B.P. 6009, F-45060 Orle´ans cedex 2, France bUMR5831-Imagerie Ge´ophysique, Universite´ de Pau, BP 1155, F-64013 Pau Cedex, France 15 Received 1 December 2002; accepted 26 May 2004 17 19 Abstract 21 The southern portion of the Upper Rhine Graben, a major oblique rift among France, Germany and Switzerland, shows a weak instrumental seismic record despite its remarkable physiographic imprint within the Northern Alpine foreland. Since traces of active 23 deformation can be found in this region and based on experience in other European areas with high seismic hazard and dense population, we searched for past earthquakes recorded in historical catalogues. Based on the fact that tectonic deformation 25 cumulates through geological time and considering that long-term effects tend to leave characteristic signatures on present-day landscape arrangement, our goal was to identify faults that could have caused the damage of recorded historical events. 27 We isolated five main earthquakes, of moderate Richter magnitude, essentially located on the E flank of the graben (as is the case with recent seismic activity). To such events, we were able to associate a specific prospective structure through the use of a procedure 29 thus far successfully employed in Southern European contexts. We concentrated on three events which showed (a) notable sensitivity to the density of the historical felt reports and (b) accordance with on-going subtle deformation pattern. Another, most relevant earthquake (M 5.5) yielded a promising match with the known deformation network in the region. 31 As a template to better constrain earthquake cycle and damage potential, historical seismicity offers an invaluable tool, since it contains a specific record, although not always unambiguous. Cross-checking such data with pertinent geological information 33 allows to devise a realistic fault geometry capable of being responsible for a specific seismic event. r 2004 Published by Elsevier Ltd. 35 57 37 1. Introduction accounted for. To this purpose, historical catalogues 59 39 contain a compilation of data describing past earth- 1.1. Detecting seismogenic sources from historical quakes. 61 41 records Such non-instrumental coverage yields a precious wealth of information, together with some uncertainties 63 43 Overall seismicity patterns are a key tool to assess (Musson, 1998; Albarello et al., 2001) which have to be ongoing deformation mechanisms in tectonically active taken into account. First of all, a catalogue only 65 45 regions. Such patterns show up from recorded seismic contains felt reports. These can be described as local event of the last 25–30 years, that is after the inception and contemporary records of damages reported either as 67 47 of permanent seismic networks. Nevertheless, to ascer- historical documents or private correspondence. Such tain the recurrence time of a given earthquake and to documentation is then translated by present-day ap- 69 49 search for a prospectiveUNCORRECTED fault potentially causative of an proaches into PROOF an objective evaluation of actual damages older event, a much longer timespan needs to be (sometimes below or beyond descriptions), expunging 71 51 style of report and balancing the true authenticity of à Corresponding author. INGV, Via di Vigna Murata 605, I-00143 various documents that describe a recognizable location. 73 53 Roma, Italy. Tel.: +39-06-51860557; fax: +39-06-51860507. E-mail address: [email protected] (U. Fracassi). 75 55 0277-3791/$ - see front matter r 2004 Published by Elsevier Ltd. doi:10.1016/j.quascirev.2004.05.009 JQSR : 1459 ARTICLE IN PRESS 2 U. Fracassi et al. / Quaternary Science Reviews ] (]]]]) ]]]–]]] 1 The ultimate goal, clearly, is to then render the entire and their descriptions into a coherent document (Boschi 57 process into a single intensity value for a given locality. et al., 2000; Albarello et al., 2001). In this study, we 3 As one moves back in time, languages, dialects, merged the former input with geomorphological and 59 documents, hand drawings etc. all become less obvious. geological data, augmented by seismic events from the 5 However, even in ideal situations, i.e. a readable and instrumental era. This allowed us to infer potential 61 self-consistent dataset of records for a given date/event, seismogenic sources for the most prominent (epicentral 7 compilation of damage isolines, needed to pin down an intensity4 VI MSK degree) historical earthquakes 63 intensity pattern, remains a complex task. This is covering the last 1200 years. (Lambert et al., 1996) 9 specifically due to two main facts: (a) local intensities between the border fault systems that flank the S portion 65 in the catalog were derived from descriptive information of the Upper Rhine Graben. 11 and (b) identifying a precise location for a specific The background data set we employed for most maps, 67 record may not be univocal. A typical example is our calculations and geomorphic evaluation is a Digital 13 study region, which underwent different dominations Elevation Model. It was obtained from the merger of the 69 across the centuries (a common feature in Europe; see French source (courtesy of BRGM) at a nominal 15 Esposito et al., 1995; Lambert and Winter, this volume) resolution of 50 m and the German one (courtesy of 71 and even hosted a local language (Alsatian), so that Universita¨t Freiburg) with a 25 m posting. Although the 17 locality names were altered back and forth from one country border corresponds to the Rhine river course, 73 tongue to another. The result commonly caused the German dataset actually covers a broad portion of 19 placenames which have some shared traits in two similar interior French territory. This resulted into a merged 75 dialects (and that could therefore be considered as a database with a 25 m resolution for most of the research 21 single locality) to eventually appear under two distinct area. 77 sites—and possibly coordinates as well. On the other 23 79 hand, since intensity values and geographical location of 1.2. Geological and geodynamic scenario of the Upper felt reports are the critical information in a historical Rhine Graben 25 catalogue, their reliability influences the overall assess- 81 ment of a prospective seismogenic source. The Rhine Graben is an oblique rift (Ziegler, 1992), 27 Furthermore, while historical records contain a 83 300 km long and 40 km wide, from Mulhouse (France) wealth of information, often not entirely exploited, they to Frankfurt (Germany), trending SSW-NNE. As per its 29 also inherit the material complexities needed to compile 85 imprint in the geodynamic arrangement of Western them. Most of the evaluation about the data reliability Europe (Fig. 1), the rift basin is hosted within the 31 remains solely with the database curators, which are 87 Vosges Mountains to the W, the Black Forest to the E, thence meant to convey data arising from felt reports the Rhenish Massif to the N and the Jura thrustbelt to 33 89 35 91 37 93 39 95 41 97 43 99 45 101 47 103 49 UNCORRECTED PROOF 105 51 107 53 109 55 Fig. 1. Main regional elements of continental Europe and bounding basins, including the southern Upper Rhine Graben (in the red box) 111 (topographic data courtesy of Scripps Institution of Oceanographyr 1987). JQSR : 1459 ARTICLE IN PRESS U. Fracassi et al. / Quaternary Science Reviews ] (]]]]) ]]]–]]] 3 1 the S (Becker, 2000). Together with the Vienna Basin, it distribution and occurrence with the landscape evolu- 57 is a prominent feature of the N Alpine Foreland (Illies tion in the Upper Rhine Graben. 3 and Greiner, 1978; Ziegler, 1992). 59 Its inception dates back to the Upper Eocene, starting 1.3. Long-term vs. short-term indicators of recent activity 5 with E–W extension (Bois, 1993), apparently in accor- 61 dance with pre-existing weak zones (Bonjer et al., 1984; The former lines of evidence indicate a seeming 7 Clauser et al., 2002). From Eocene until Aquitanian seismotectonic inconsistency, in as much as a certain 63 time, the graben was mainly fed by shales and seismic activity on the SE shoulder of the Upper Rhine 9 sandstones; extensive evaporites deposited in the S Graben identifies an ongoing deformation in the vicinity 65 portion (Bois, 1993; Sissingh, 1998; Derer, 2003). Since of the normal border fault. Conversely, the sedimentary 11 the Middle Pliocene, the tensional graben twisted into a pattern discussed by Menillet (1995) and depicted in Fig. 67 broader overall shear zone, predominated by a strike- 2a highlights a depositionary scenario predominated by 13 slip kinematics (Ahorner, 1975) and by the typical 3D sedimentation currently detectable only along the W 69 oblique rift arrangement (McClay et al., 2002). The margin, with a major river system flowing towards NE 15 graben cuts through the Lalaye–Lubine–Baden Baden (Bru¨stle, 2002). Whatever present-day feeding from the 71 crustal discontinuity, a remnant of the Variscan orogen; Black Forest is either concealed by the Rhine thalweg, 17 the accommodation zone of the Rhine Graben devel- as borehole data indicate (Bru¨stle, 2002), or smoothed 73 oped across this major oblique fracture system (Brun et away towards N by the Rhine itself. The river’s 19 al., 1992; Mayer et al., 1997). sweeping course likely competed with the vosgian 75 The shape of the Oligocene depositional top indicates alluvium, acting either as a mere E boundary to the 21 the quantitative difference in subsidence and its geome- fan system or trimming its distal aprons, as often 77 try along the Rhine Graben (Brun et al., 1992).