Geochemical and temporal constraints on mantle heterogeneity and climate/volcano connections using sediment-hosted glasses and basalt olivine

Abstract

Oceanic crust is continuously recycled into the mantle at subduction zones, generating chemical heterogeneity that is ultimately expressed in oceanic basalts. A central question in mantle geochemistry is how this recycled material is stored and how it contributes to mantle melting. One class of models proposes that recycled oceanic crust persists as a distinct lithology, such as pyroxenite, which undergoes extensive melting and contributes directly to basalt genesis, forming a veined mantle. An alternative model suggests that recycled crust instead produces small-degree melts that metasomatize the surrounding peridotitic mantle, creating a “fertilized” source that subsequently melts to generate basalts. Distinguishing between these end member scenarios is critical for understanding the origin of ocean island basalts (OIB), mid-ocean ridge basalts (MORB), and mantle heterogeneity more broadly. This thesis addresses this problem through geochemical investigations of basalts and olivine from Hawaii, volcanic glasses from Iceland, and a time-resolved study of mid-ocean ridge magmatism using sediment-hosted glasses (SHGs) from the Reykjanes Ridge. In addition to constraining mantle source lithology, it establishes SHGs as a new archive of ocean ridge magmatism by developing the analytical methods required to measure them and by constructing the first high-resolution SHG-based geochemical time series. Chapter 1 investigates the origin of Ni and Mn variations in olivine phenocrysts from Hawaiian basalts and MORB. High Ni contents in Hawaiian olivine have often been interpreted as evidence for pyroxenite-derived melts. By integrating olivine compositions with coexisting liquid data and quantitative polybaric melting models, this study shows that neither pyroxenite melting nor simple peridotite melting alone can explain the observed systematics. Instead, the data are best explained by the addition of small fractions of low-degree melts derived from recycled eclogitic crust to a peridotitic mantle source, followed by low extents of melting at high pressure. These results demonstrate that OIB geochemical signatures can be reproduced without invoking a distinct pyroxenite lithology, supporting a metasomatized peridotite model. This model is further evaluated using Icelandic basaltic glass data, motivating the second part of the thesis. Temporal variability in ridge magmatism is difficult to resolve using conventional seafloor sampling, but SHGs deposited in nearby marine sediments provide a unique archive of erupted magmas spanning up to hundreds of thousands of years. To exploit this archive, approximately 3,000 SHGs from sediment cores near the southern Reykjanes Ridge were analyzed for major and trace elements using newly developed high-throughput analytical methods. These samples define a wide compositional range derived from both the local ridge system and nearby Icelandic volcanic centers. Chapter 2 develops a quantitative framework to distinguish these diverse SHG compositions. Using multivariate clustering of elemental abundances and ratios, combined with statistical mapping, the samples are separated into geochemical groups representing the ridge axis, off-axis volcanism, and Icelandic eruptions. This analysis also demonstrates the limitations of purely unsupervised approaches and the need to integrate statistical methods with geochemical constraints. Building on this framework, Chapter 3 reconstructs the first high-resolution geochemical time series (~130 kyr) of mid-ocean ridge magmatism from ridge-derived SHGs. Variations in incompatible element concentrations and MgO correlate with glacial–interglacial sea-level changes, with increased enrichment during the last deglaciation. These patterns are best explained by reduced mantle melting during periods of rapid sea-level rise, with short lag times indicating efficient melt transport. This provides direct geochemical evidence that climate-driven sea-level changes can modulate mid-ocean ridge magmatism. Chapter 4 focuses on SHGs derived from Icelandic volcanic systems to investigate mantle source characteristics and temporal variations in magma differentiation over the past ~80 kyr. The data define distinct geochemical populations corresponding to different volcanic systems. Most show stable differentiation behavior through time, although some exhibit temporal variability. Parental magmas for historic samples appear to be derived by slightly larger extents of melting, as would be predicted from the effects of reduced ice loading. Trace element systematics indicate that the Icelandic mantle source is best described as peridotite metasomatized by low-degree melts of recycled crust, consistent with Chapter 1. Highly incompatible element ratios (e.g., Nb/U, Ba/Th, K/U) suggest a minor additional metasomatic component, possibly related to carbonatite enrichment. These results further support fertilized peridotite as the primary source of mantle heterogeneity beneath Iceland. Chapter 5 presents the key methodological contribution: the development of a rapid, accurate LA-ICP-MS protocol for simultaneous major and trace element analysis of silicate materials. This approach enables the large, high-quality SHG datasets required for this study and establishes a framework for future work using similar archives. Taken together, these results show that mantle heterogeneity beneath both ocean islands and mid-ocean ridges is best explained by metasomatized peridotite formed through interaction with low-degree melts of recycled oceanic crust, rather than by distinct pyroxenite lithologies. More broadly, this work establishes sediment-hosted glasses as a powerful new tool for reconstructing time-resolved magmatic processes and demonstrates that climatic forcing, through sea-level change, exerts a direct and measurable influence on mantle melting and magma generation at mid-ocean ridges.