Cooling history and emplacement dynamics within rubbly lava flows, southern Deccan Traps: insights from textural variations and crystal size distributions

Abstract

We analyse two representative rubbly pāhoehoe lavas (F3 and F5) from drill cores at Tural-Rajwadi, southwest of Koyna, in the southern Deccan Traps. Low vesicle deformation (0.1 to 0.4) indicates that both lavas ultimately cooled under a low-stress regime. The crystal size distributions (CSDs) of most samples from F5 (especially those from within the core) are not linear but instead show kinks. These kinks are attributed to a rise in plagioclase nucleation due to degassing following the brecciation of the crust. Since it is difficult to constrain cooling time for ancient lava flows, we used the products of nucleation rates (Jt, 1.64 × 10–8 to 1.45 × 10–5 μm−3) and growth rates (Gt, 2.1 to 156 μm) with time. When compared with natural analogues as well as experimental results for basalt crystallisation, these values suggest a much faster lava cooling rate (~ 1 to 7℃/hr) than a conductive cooling model (≤ 0.1 ℃/hr). The CSDs for F3 fan with depth suggesting that the lava flow might represent local accumulation (ponding?) in a transitional lava flow field. CSDs for F5 show little variation with depth, with the exception of kinks for samples from the lower crust and core. The relatively higher number density of plagioclase microcrysts in our rubbly pāhoehoe (F5) and their CSD patterns are similar to those measured for transitional lavas from Hawaii. The vesicle data and CSDs indicate that brittle deformation was the primary mode of transition within these lavas. Identifying occurrence of thick ponded lavas within vertical stacks of rubbly pāhoehoe flows in the upper stratigraphic levels of the Deccan Traps are critically important as they demonstrate complex cooling styles, crystallisation histories, and emplacement dynamics. Transitional lavas such as rubbly pāhoehoe are important components of large CFB provinces such as the Deccan Traps and constitute nearly 46 to 85% of all lava types. Modelling of continental flood basalt provinces should therefore account for these diversities within lavas, and any oversimplified version using end-member morphotypes is unrealistic and untenable.