GNGTS 2016 - Atti del 35° Convegno Nazionale

306 GNGTS 2016 S essione 2.1 According to the International Seismological Centre (ISC) worldwide bulletin (ISC, 2012) for Cameroon from 1900 to 2010 the strongest event occurred in September 1945 at the border with Congo and a strong sequence happened in January 1969 in the area of the Mbam et Djerem National Park with an event of magnitude 5.8 followed one month later by a strong aftershock of magnitude 5.5. The Centennial Catalogue (Engdahl and Villaseñor, 2002) reports only the event of September 23, 1974 in Gabon of magnitude around 6. The Global Earthquake Catalogue for Stable Continental Regions (Schulte, 2005) reports the quake of March 26, 1911 in the area of Lolodorf (Ambraseys and Adams, 1991) with a magnitude of 5.8. According to Ambraseys and Adams (1986), two additional quakes hit the coastal area of Cameroon in 1903 (magnitude 4.4) and 1908 (magnitude 4.3). Scientific studies on the Kribi area (Nfomou et al. , 2004) accounts for sixteen additional earthquakes in southern Cameroon during the last hundred years and the last one recorded in July 2002 (magnitude 3.2). The exclusion of the dependent events (6 over all) was done according to the Gardner and Knopoff (1974) approach. The final catalogue consists of 58 earthquakes:10-1400 km the range of distances and magnitude between 2.7 and 6.1. For delineating seismogenic sources an adequate kinematic model is needed, capable of explaining the recent evolution and the seismicity of the whole area. Western Africa is a stable continental area, mostly a very ancient plateau and it displays few clearly recognised active tectonic features and a highly segmented margin by many fracture zones and volcanic lines. Merging together the data from the earthquake catalogue and the available tectonic information enabled us to identify the seismotectonic model for the study region. The model has 4 zones and a background zone collecting the low seismicity not linked to any fault. For the zonation based on actual faults we considered all the faults as active faults. The fault activity was modelled with a geodetic approach based on slip rate values estimated for each fault from GPS data modelling and the geometry of the fault rupture. According to Kreemer et al. (2003; available at http://gsrm.unavco.org/ ), the regional slip velocity is taken as 25 mm/ year toward NE (51°). We considered a possible segmentation for the major faults. We decided to apply two GMPEs of global applicability: Akkar and Bommer (2010; A&B hereafter) and Cauzzi and Faccioli (2008; C&F hereafter). A logic tree (Fig. 1) with 4 branches was constructed for the bedrock hazard maps; it consists of two source models: a model with wide SZs, fed by a GR magnitude model with a statistical computation for the maximum magnitude, and a model with faults, associated with a characteristic earthquake seismicity model (Schwartz and Coppersmith, 1984) with a tectonic estimation of the maximum magnitude and two aforementioned GMPEs. The mean results are aggregated by the weighting scheme (Fig. 1) for the different branches of the logic tree. An updated 2012 version of the computer code Crisis (Ordaz et al. , 2013) was applied. The main results of the PSHA are shown in terms of seismic hazard curves and uniform hazard response spectra (UHRS) for rock. In addition to the seismic hazard curve six return periods (RPs) have been considered: 95, 475, 712, 975, 2475, and 4975 years (Tab. 1), that correspond to the following exceedance probability in 50 years: 41%, 10%, 6.78%, 5%, 2%, and 1%, respectively. The 95- and 712-year return periods correspond in the Eurocode 8 to the two limit states of damage limitation requirement and life preservation, respectively, when industrial buildings are considered. They could be, therefore, of particular interest for an industrial structure as the Kribi port complex (Fig. 1). Tab. 1 – Main PSHA results for Kribi site. PGA [g] RP [years] 475 712 975 2475 4975 BEDROCK 0.07 0.09 0.10 0.16 0.22