Mantle plume and rift-related volcanism during the evolution of the Rio Grande Rise

Oceanic plateaus are extensive elevated regions on the ocean floor, formed primarily through excess magmatism. Their origins are often attributed to mantle plumes, where upwelling hot mantle material leads to significant crustal formation, sometimes exceeding 35 km in thickness. However, some plateaus represent fragments of continental crust, resulting from continental rifting and subsequent isolation within oceanic basins. The Rio Grande Rise (RGR) and the Walvis Ridge in the South Atlantic Ocean are significant bathymetric features, previously interpreted as Large Igneous Provinces related to the Tristan-Gough mantle plume. Existing data, including trace element and isotopic analyses, and 40Ar-39Ar ages, suggest a joint formation of the Walvis Ridge and western RGR around 120-80 Ma, during plume-ridge interaction. Between 80 and 60 Ma, continued volcanic activity separated the two features, with the formation of the eastern RGR. Subsequent ridge jumps isolated the RGR around 60 Ma, though plume-related volcanism persisted on the African plate. A later volcanic event at ~46 Ma is recorded in the western RGR, characterized by a distinct geochemical signature. The Rio Grande Rise is intersected by the Cruzeiro do Sul Rift (CdSR), a significant tectonic feature also impacting the Brazilian margin. The Jean Charcot Seamount Chain (JCSC) extends northwestward from the CdSR. Recent studies have proposed an alternative hypothesis, suggesting that the western RGR is a continental fragment, a 'microcontinent,' based on geophysical data (negative Bouguer anomaly) and the discovery of Proterozoic age gneiss, granite, and pegmatite. This study aims to resolve this debate by presenting new geochemical data from volcanic rocks across the RGR and JCSC, comparing them to data from Deep Sea Drilling Project (DSDP) Site 516 (a well-studied area of the RGR) to better understand the formation and evolution of these structures and their relationship to the Tristan-Gough plume and the CdSR. The research also investigates evidence for the existence of remnants of continental crust beneath the RGR.

Several models attempt to explain the formation of the Rio Grande Rise. Early models emphasized its origin as a Large Igneous Province (LIP) resulting from the Tristan-Gough mantle plume's activity, with the plume situated near the Mid-Atlantic Ridge during the formation of the Walvis Ridge and western RGR. This model is supported by geochemical and geochronological data from the Walvis Ridge and DSDP Site 516, showing isotopic similarities and ages consistent with plume-ridge interaction. However, alternative hypotheses have emerged, primarily proposing the western RGR as a microcontinent detached from the South American margin and subsequently covered by younger basaltic lavas. These hypotheses cite the presence of a negative Bouguer anomaly and the discovery of Proterozoic age continental rocks dredged from the western RGR as supporting evidence. This study aims to reconcile these contrasting hypotheses by providing comprehensive geochemical data from newly sampled regions of the RGR and related structures.

Volcanic rock samples were collected during research cruise MSM-82 from the eastern and western RGR and the JCSC using dredging techniques. Samples were selected based on minimal alteration, and the altered rims and manganese crusts were carefully removed before analysis. Major and trace element analyses were performed using X-ray fluorescence (XRF) spectrometry for major elements and inductively coupled plasma mass spectrometry (ICP-MS) for trace elements. A total of 83 samples were collected, but only 56 (loss on ignition <5 wt.%) were used for the major geochemical interpretations to minimize the effects of seawater alteration. Two samples with 5.3 and 5.8 wt% loss on ignition were also used for reference as they appeared relatively fresh. The study focused on fluid-immobile elements, including Ti, Y, Zr, Nb, Hf, Th, and rare earth elements (REE), which are considered resistant to seawater alteration. Several classification diagrams, using ratios of incompatible trace elements, were employed to discriminate between different volcanic rock types and tectonic settings. The authors also used primitive mantle-normalized multi-element variation diagrams to illustrate compositional differences between samples from various locations within the RGR and JCSC. The REEBOX Pro (v.1) melting algorithm was applied to estimate the melting conditions of both tholeiitic and alkaline magmatic phases. This model calculates trace element compositions of melts over a range of pressures and temperatures, considering different potential mantle sources (depleted MORB mantle and primitive mantle) for the tholeiitic and alkaline lavas. The model allowed for the estimation of the degree of melting and depth of melt generation for different samples, and provided insight into the thermal conditions and crustal thickness during the formation of the RGR.

The geochemical analysis revealed distinct compositional differences between the eastern and western RGR. Eastern RGR and older parts of the western RGR are characterized by tholeiitic basalts with moderately enriched trace element compositions, consistent with formation at or near a plume-influenced spreading center. These samples show geochemical affinities with the Walvis Ridge and the DSDP Site 516 samples. In contrast, younger, alkaline lavas from the western RGR and JCSC formed through lower degrees of partial melting beneath thicker lithosphere, suggesting an intraplate setting, potentially related to rifting. The alkaline lavas show geochemical similarities to the Tristan da Cunha and Gough Island lavas. The study found no geochemical evidence for significant continental crustal assimilation in any of the RGR lavas. Nb/Zr and Nb/Th ratios show no systematic decrease with decreasing MgO, indicating that fractional crystallization did not significantly influence these ratios. Melting modeling, using the REEBOX Pro algorithm, indicated that the tholeiitic lavas formed through moderate degrees of partial melting (2-5%) at relatively low pressures (<2.7 GPa), consistent with a plume-ridge interaction scenario. In contrast, the alkaline lavas resulted from lower degrees of melting (0.5-2%) at higher pressures (>2.9 GPa), indicating melting beneath a thicker lithosphere. The modeled mantle potential temperatures for both tholeiitic and alkaline lavas were higher than average asthenosphere temperatures, further suggesting a role for mantle plumes. The transition from tholeiitic to alkaline volcanism spanned tens of millions of years, unlike typical late-stage alkaline volcanism associated with mantle plumes. The geochemical similarities between alkaline lavas from the western RGR and the JCSC suggest a possible connection between these features, possibly linked to Eocene rifting and the formation of the CdSR.

The findings demonstrate that the RGR is not a simple, homogenous structure but rather a complex volcanic plateau with a protracted history of magmatic activity. The early formation of the RGR involved plume-ridge interaction, resulting in the emplacement of tholeiitic basalts. Subsequent rifting, as evidenced by the CdSR, led to the formation of the alkaline lavas in the western RGR and possibly the JCSC. The lack of evidence for crustal contamination supports the predominantly volcanic origin of the RGR, contrasting with previous hypotheses suggesting a significant continental component. The age difference and geochemical variations between the tholeiitic and alkaline lavas suggest different sources and degrees of melting, indicating heterogeneous mantle source regions beneath the western South Atlantic. The similarity of the alkaline lavas to those from the Gough and Tristan da Cunha islands points to a connection with the Tristan-Gough hotspot track. The temporal relationship between rifting, mantle plume activity, and the onset of alkaline volcanism suggests a complex interplay of tectonic and magmatic processes in shaping the RGR.

This study provides strong evidence for a predominantly volcanic origin for the Rio Grande Rise, formed through a complex interaction of mantle plume activity and lithospheric rifting. The distinct geochemical signatures of tholeiitic and alkaline lavas reflect distinct stages in the RGR's evolution, with plume-ridge interaction characterizing the earlier stage and intraplate rifting associated with the later alkaline stage. The lack of evidence for crustal contamination refutes previous hypotheses of a significant continental crustal component. Future research could focus on more precise geochronological dating of the alkaline lavas to further refine the timing of rifting events and their relationship to magmatic activity. Integrating geophysical data, such as seismic tomography, with the geochemical data could also improve our understanding of the mantle structure and its role in shaping the RGR's evolution.

The study's reliance on dredged samples may not fully capture the entire compositional range of volcanic rocks within the RGR. The effects of alteration, even in carefully selected samples, may still influence the interpretation of certain geochemical parameters. The melting models require assumptions about the mantle source composition, and uncertainties in these assumptions may affect the accuracy of the derived melting parameters. Furthermore, the lack of direct dating of the JCSC lavas limits a detailed comparison with the western RGR alkaline samples.