Upper Mantle Temperature of Hotspots and Ridges and Correlation with Geochemical Signals

The temperature of hotspots compared to those of mid-ocean ridges has been the subject of debate. We reexamine the differences in temperature and geochemical signals between hotspots and ridges by attempting to obtain temperature directly from seismic tomography. Previous studies of the potential temperature based on petrological models (Putirka, 2005) and transition zone thickness (Courtier et al., 2007) suggested the following decreasing order of temperature: hotspots, ridge-influenced hotspots, and ridges. Previous studies had only examined a few locations as catalogues and measurements were not as complete. Jackson et al. found compelling relationships between the signals of geochemical enrichment at hotspots with tomography(2017) and distance to LLSVPs(2018). Dalton et al.(2014) converted seismic velocity and ridge depth to temperature for ridges, and also found correlations with major geochemical signals. Their work also showed a strong hemispherical difference in the potential temperature of ridges. Here, we use HeFESTo (Stixrude and Lithgow-Bertelloni, 2011) to convert velocity to temperature. We assume a pyrolytic, compositionally homogenous upper mantle based on the DMM composition of Workman and Hart (2005). We use several global and regional S-wave tomographic models to infer the temperature beneath hotspots and ridges in the upper mantle for the global hotspot catalog of Jackson et al. (2018) and the ridge segments of Gale et al. (2013). Our results show that: 1) not all hotspots are hotter than ridges, but hotspots with higher ³He/⁴He signal tend to be hotter; 2) ridges close to hotspots are mostly hotter than other ridges; 3) ridges are hotter than the 1600 K adiabat. Using cluster analysis, we also find distinct regional differences. Pacific hotspots and ridges are overall hotter than African and Atlantic ones, consistent with Dalton et al.(2014), perhaps indicative of differences in nature between the African and Pacific LLSVPs. The fact that our reference 1600 K adiabat is colder than the median temperature of ridges far from hotspots implies that MORB may melt at higher temperatures, as suggested by Sarafian et al. (2017). Cluster analysis techniques are also developed to find possible plumes paths from the surface expression of hotspots to LLSVPs.