THE USE OF GPR FOR THE DETECTION OF NON-HOMOGENEITIES IN THE RENO RIVER EMBANKMENTS (NORTH-EASTERN ITALY) Biavati G. (1) , Ghirotti M. (1) , Mazzini E. (2) , Mori G. (1) , Todini E. (1) (1) Department of Earth and Geoenvironmental Sciences, University of Bologna, Italy. (2) Regione Emilia-Romagna Regional Authority, Basin Technical Survey, Italy. (giulia. biavati@unibo.it / Phone +39 051 2094565) RÉSUMÉ L’utilisation de techniques non-invasives lors de l’analyse des berges artificielles des fleuves est essentielle pour la sauvegarde du fleuve lui-même. Récemment, ce thème a fait l’objet d’une attention particulière lors de l’utilisation de méthodes géophysiques. Contrairement aux autres méthodes, le GPR offre la possibilité d’étudier des zones étendues sur un laps de temps relativement court. Lors de l’analyse de berges, il permet de détecter des anomalies pouvant être étudiées plus tard en détails. La sécurité hydraulique du fleuve Reno, l’un des plus importants du Nord-Est de l’Italie, constitue une priorité pour l’administration de la région Émilie-Romagne. Au cours des dernières décennies, de nombreux événements hydro-météorologiques ont provoqué des dommages structurels dans les berges du fleuve Reno et ses affluents. Le potentiel du GPR a été testé à différentes fins : détection de la stratigraphie des berges, étude du développement de cavités creusées par des animaux, localisation de conduits inconnus et de zones réparées. Les premiers résultats montrent que le GPR est, en termes de rapidité d’acquisition et de précision, une technique efficace pour la détection d’anomalies telles que les zones réparées, les niveaux de construction, les éléments cachés et les principaux horizons stratigraphiques peu profonds. ABSTRACT The use of non-invasive techniques in the monitoring of man-made river embankments is of primary importance for the safety of the embankment itself. Recently, a noticeable attention for this topic has been addressed to many geophysical methods. The GPR method respect to the others gives the possibility to investigate spread areas in a relative short time; in the study of river embankments, it allows to detect anomalies that can be thoroughly studied later on. The hydraulic safety of the Reno River, one of the main rivers in North-Eastern Italy, is of primary importance for the Emilia-Romagna regional administration. In the last decades, several hydro-meteorological events caused structural failures of the embankments of the Reno River and its tributaries. The GPR capability has been tested in several test areas, along more than 15 km on artificial embankments for several issues: detection of river embankment stratigraphy, study of the development of cavities produced by animals, localization of unknown pipelines and repaired areas. Preliminary results show that GPR is a suitable technique for the detection of anomalies like levels of construction, recently repaired areas, hidden objects, important shallow stratigraphic horizons. 1. INTRODUCTION The problem of the safety of river embankments can be easily underestimated, particularly if a relative dry period persists for a long time. The Reno river together with its tributaries covers, in the alluvial plain area, a length extension of more than 500 km. Some crisis events happened in the past highlighted how most of problems that can interest embankments occur along discontinuities hardly detectable with traditional techniques. Piping and seepage problems along fractures, cavities, differently compacted and settlemented areas, cannot be easily recognised through destructive methods which provide precise but punctual information. Non-invasive techniques, such as geophysical methods, result largely diffused for this topic and in constant development. In particular, seismic refraction and MASW (Multichannel Analysis of Surface Waves) techniques are the most suitable both for detecting different compacted areas and for estimating geotechnical properties of the materials; however, they are ineffective for identifying the presence of cavities and burrows (Morris, 2005). Electric tomography methods are able to resolve almost all the problems mentioned above with good resolution, mostly as regard to the structure and soil moisture content; but they require long time acquisition, because the spacing between the electrodes should be equal to or below 1 m. Innovative electromagnetic method as GMS (Geophysical Monitoring System), employing the GEM-2 (Boukalová and Beneš, 2007), was tested and verified in European Projects; it carries out basic assessment of the dikes condition and their material composition in large areas with quick and not highly money- consuming measurement. Looking for a higher resolution for the detection of non-homogeneities, the attention has been turned to alternative methods such as ground-penetrating radar (GPR). After an experimental period concerning mostly archaeological and urban purposes (Al-Quadi and Lahouar, 2005; Loizos and Plati, 2007), interest in GPR survey for geological issues increased considerably in the 1990s. As an example, the GPR has unquestionable advantages, respect to other techniques for the estimation of scree thickness of hardly accessible mountain areas (Otto and Sass, 2005). Moreover, GPR capability in the determination of buried geological structures makes this technique a powerful tool in sedimentology (Smith and Jol, 1995); even if the extraction of meaningful information on In : J. Locat, D. Perret, D. Turmel, D. Demers et S. Leroueil, (2008). Comptes rendus de la 4e Conférence canadienne sur les géorisques: des causes à la gestion. Proceedings of the 4th Canadian Conference on Geohazards : From Causes to Management. Presse de l’Université Laval, Québec, 594 p.