The Central Atlantic Magmatic Province (CAMP) in Morocco

The Central Atlantic Magmatic Province (CAMP) is a large igneous province (LIP) composed of basic dykes, sills, layered intrusions and lava flows emplaced before Pangea break-up and currently distributed on the four continents surrounding the Atlantic Ocean. One of the oldest, best preserved and most complete sub-provinces of the CAMP is located in Morocco. Geochemical, geochronologic, petrographic and magnetostratigraphic data obtained in previous studies allowed identification of four strato-chemical magmatic units, i.e. the Lower, Intermediate, Upper and Recurrent units. For this study, we completed a detailed sampling of the CAMP in Morocco, from the Anti Atlas in the south to the Meseta in the north. We provide a complete mineralogical, petrologic (major and trace elements on whole-rocks and minerals), geochronologic (40Ar/39Ar and U-Pb ages) and geochemical set of data (including Sr-Nd-Pb-Os isotope systematics) for basaltic and basaltic-andesitic lava flow piles and for their presumed feeder dykes and sills. Combined with field observations, these data suggest a very rapid (<0.3Ma) emplacement of over 95% of the preserved magmatic rocks. In particular, new and previously published data for the Lower to Upper unit samples yielded indistinguishable 40Ar/39Ar (mean ag e= 201*2 +-0.8Ma) and U-Pb ages (201*57=-0.04Ma), suggesting emplacement coincident with the main phase of the end-Triassic biotic turnover (c.201*5 to 201*3Ma). Eruptions are suggested to have been pulsed with rates in excess of 10km3/year during five main volcanic pulses, each pulse possibly lasting only a few centuries. Such high eruption rates reinforce the likelihood that CAMP magmatism triggered the end-Triassic climate change and mass extinction. Only the Recurrent unitmay have been younger but by no more than 1Ma. Whole-rock and mineral geochemistry constrain the petrogenesis of the CAMP basalts. The Moroccan magmas evolved in mid-crustal reservoirs (7-20 km deep) where most of the differentiation occurred. However, a previous stage of crystallization probably occurred at even greater depths. The four units cannot be linked by closedsystem fractional crystallization processes, but require distinct parentalmagmas and/or distinct crustal assimilation processes. EC-AFC modeling shows that limited crustal assimilation (maximum c.5- 8% assimilation of e.g. Eburnean or Pan-African granites) could explain some, but not all the observed geochemical variations. Intermediate unit magmas are apparently the most contaminated and may have been derived from parental magmas similar to the Upper basalts (as attested by indistinguishable trace element contents in the augites analysed for these units). Chemical differences between Central High Atlas and Middle Atlas samples in the Intermediate unit could be explained by distinct crustal contaminants (lower crustal rocks or Pan-African granites for the former and Eburnean granites for the latter). The CAMP units in Morocco are likely derived from 5-10% melting of enriched peridotite sources. The differences observed in REE ratios for the four units are attributed to variations in both source mineralogy and melting degree. In particular, the Lower basalts require a garnet peridotite source, while the Upper basalts were probably formed from a shallower melting region straddling the garnet-spinel transition. Recurrent basalts instead are relatively shallow-level melts generated mainly from spinel peridotites. Sr-Nd-Pb-Os isotopic ratios in the CAMP units from Morocco are similar to those of other CAMP sub-provinces and suggest a significant enrichment of the mantle-source regions by subducted crustal components. The enriched signature is attributed to involvement of about 5-10% recycled crustal materials introduced into an amb