Stable carbon and oxygen isotope ratios of carbonates of ODP Site 103-639

A Tithonian sequence of shallow-water limestones, intercalated with siliciclastics and overlain by dolomite, was recovered during drilling at ODP Site 639 on the edge of a tilted fault block. The carbonates were strongly affected by fracturing, dolomitization, dedolomitization, and compaction. The chronology and nature of the fractures, fracture infilling, and diagenesis of the host rock are established and correlated for both the limestone and the dolomite. A first phase of dolomitization affected limestone that was already, at least partially, indurated. In the limestone unit, fractures were filled by calcite and dolomite; most of the dolomite was recrystallized into calcite, except for the upper part. In the dolomitic unit, the first-formed dolomite was progressively recrystallized into saddle dolomite, as fractures were simultaneously activated. The dolomitic textures become less magnesian (the molar ratio mMg/mCa goes from 1.04-0.98 to 0.80), and the d18O (PDB) ranges from -10 per mil to -8 per mil. The varying pores and fissures are either cemented by a calcic saddle dolomite (mMg/mCa ranging from 0.95 to 0.80) or filled with diverse internal sediments of detrital calcic dolomite, consisting of detrital dolomite silt (d18O from -9 per mil to -7 per mil) and laminated yellow filling (with different d18O values that range from -4 per mil to +3 per mil). These internal sediments clearly contain elements of the host rock and fragments of saddle crystals. They are covered by marls with calpionellids of early Valanginian age, which permits dating of most of the diagenetic phases as pre-Valanginian. The dolomitization appears to be related to fracturing resulting from extensional tectonics; it is also partially related to an erosional episode. Two models of dolomitization can be proposed from the petrographic characteristics and isotopic data. Early replacement of aragonite bioclasts by sparite, dissolution linked to dolomitization, and negative d18O values of dolomite suggest a freshwater influence and 'mixing zone' model. On the other hand, the significant presence of saddle dolomite and repeated negative d18O values suggest a temperature effect; because we can dismiss deep burial, hydrothermal formation of dolomite would be the most probable model. For both of these hypotheses, the vadose filling of cavities and fractures by silt suggests emersion, and the different, and even positive, d18O values of the last-formed yellow internal sediment could suggest dolomitization of the top of the sequence under saline to hypersaline conditions. Fracturing resulting in the reopening of porosity and the draining of dolomitizing fluids was linked to extensional tectonics prior to the tilting of the block. These features indicate an earlier beginning to the rifting of the Iberian margin than previously known. Dolomitization, emersion, and erosion correspond to eustatic sea-level lowering at the Berriasian/Valanginian boundary. Diagenesis, rather than sedimentation, seems to mark this global event and to provide a record of the regional tectonic history