ELSEVIER Lithos 37 (1996) 223-243 LITHOS 0 Shifts in the source of the Paran6 basalts through time F. Garland ‘, S. Turner, C. Hawkesworth zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR Depurtment oj’Earth Sciences, The Open University, Milton Keynes, MK7 6AA, UK Received 18 October 1994; accepted 25 July 1995 Abstract Magmatism occurred intermittently in the Parani region of South America between 240-50 Ma, with the peak of magmatic activity between 138-126 Ma, comprising the Paran continental flood basalts (CFB). Detailed chemical subdivision of the ParanSl basalts into six units allowed stratigraphic correlations to be made, implying a northward migration of magmatism following the rifting of the South Atlantic. However, “OAr-3gAr ages do not uphold the chronostratigraphic nature of such magmatic units, instead they indicate a southeastwards migration. Major elements (extrapolated to 8 wt.% MgO) are used to investigate changes in magma type with time, and are interpreted in terms of parameters such as pressure, depth of melting and the nature of the source region (fertile/refractory and water content>. The oldest Parana unit originated by melting at the greatest depths (90-120 km), shifting to shallow levels (30-60 km) for the younger units. Accompanying this decrease in pressure, the source material undergoing partial melting became progressively more refractory in major elements and also more hydrous. This hydrous metasomatism occurred after major element depletion of the lithosphere at shallow levels. The data is consistent with conductive heating of a stratified source region, inferred to lie within the lithospheric mantle, which is relatively fertile but anhydrous at depth, and also appears to be younger in age than the relatively major element refractory, hydrous lithosphere at shallower depths. Thus the interaction of asthenosphere and lithosphere can be determined within an integrated age, tectonic and thermal framework. 1. Introduction Continental flood basalts (CFB) are voluminous basalt provinces (often > 1 X lo6 km3), which are surface expressions of thermal anomalies in the up- per mantle, and as such can be used to model asthenosphere-lithosphere interaction. The majority of CFB are associated with continental rifting, and much debate has been engendered as to whether active or passive rifting (plume- or extension-driven) is the cause of CFB magmatism. White and McKen- 1 Fax: 0908 655151 E-mail: fe_garland@ou.ac.uk. zie (1989) describe a passive rifting model in which extension triggers magmatism in areas of high man- tle temperatures due to decompressive upwelling of asthenospheric mantle. Hooper (1990) has argued that, in the case of the Deccan and Columbia River CFB, magmatic activity precedes extension, there- fore at least in some areas plumes initiate magma- tism, which is then followed by extension. The im- portance of plumes in CFB has long been noted, and most provinces can be traced to a presently active hotspot (e.g. Tristan da Cunha-ParanB; RCunion- Deccan; Iceland-North Atlantic Tertiary). The role that the plume plays in terms of a source region for the CFB is more ambiguous, and again much debate 0024-4937/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0024-4937(95)00038-O