EXPLORING THE NORTHERN EXTENSION OF THE REUNION HOTSPOT THROUGH GEOCHEMICAL ANALYSIS

Geochemical modeling of intrusive samples from the Bela ophiolite, in Pakistan, was done in order to determine if a genetic link existed with the Reunion hotspot track. This would add to the new northern extension of the Reunion hotspot track beyond the Deccan Traps in India, and would further our understanding of the collision of the Eurasian and Indian plates. Petrographic analyses and geochemical data were used in order to help characterize the samples. The geochemical data was obtained through XRF and ICPMS analysis.

The samples can be broken into three groups based on major and trace element affinities. All samples have low K, <1.1 wt%, with tholeiitic to calc alkaline affinities, and range from 50-77 wt% SiO₂. Average LOI for the data is 4.65 wt%. The mafic samples have continental signature based on Zr/Y-Zr ratios and enrichment in large ion lithophile elements (LILE).

Group 1 represents the felsic samples with 53.7-76.6 wt% SiO₂. These samples appear to have been generated in a volcanic arc environment based on Nb-Y ratios, with typically arc-like Nb depletion. Hence they may not be related to the Reunion Hotspot, but instead be related to the ophiolite suites they intrude.

Group 2 and Group 3 samples are mafic with 49.3-52.2 wt% SiO₂. Group 2 samples have LILE concentrations of 30-50 x chondrite and high field strength element (HFSE) concentrations from 10-20 x chondrite, with almost parallel element variations. Group 3 samples have LILE from 30-200 x chondrite and HFSE at 10-20 x chondrite, with greater enrichment in LILE relative to HFSE. Both Group 2 and Group 3 magmas were generated by minor amounts of partial melting of an enriched MORB source.

Preliminary modeling of the samples has shown that both perfect fractional crystallization (PFC) and assimilation-fractional crystallization (AFC) processes can not accurately replicate all of the element assemblages. In some instances either PFC or AFC came close to approximating observed magma evolution. For example in Group 2 a reasonable fit could be made using AFC modeling though this entailed going from a higher to a lower silica sample which can be explained by almost pure assimilation, with little or no fractionation. As such it may be possible to generate the observed chemistries through pure assimilation, once the correct assimilant chemistry has been identified.