The distribution of chemical heterogeneity beneath hotspots provides important constraints on the source of magmatism and mantle convection. Chemical heterogeneities in the upper mantle produce discontinuities that can be interrogated seismically. One such discontinuity, the X-discontinuity (230-350 km depth), is observed intermittently across the globe, with multiple possible causal mechanisms. However, to better understand the cause of the X-discontinuity and its relationship to mantle upwellings, we require global short wavelength observations, previously unresolved by reflected phases with broad upper mantle Fresnel zones. Discontinuities resulting from seismically observable impedance contrasts can be targeted using high frequency P wave receiver functions (RFs) possessing short wavelength lateral sensitivity (~100 km) to structures almost directly beneath the station. Thus, below well-instrumented hotspots we can investigate the properties of the X-discontinuity and constrain the causal mechanisms in these locations.

Using P-to-s converted teleseismic phases between 30-90°, we stack RFs in the depth and time-slowness domains at 28 hotspots and 6 cratons globally. This reveals 15 robust X-discontinuity observations beneath hotspots between 244-344 km depth. A further 10 potential observations are present beneath hotspots, and 6 null observations beneath cratons. Two causal mechanisms remain possible at the elevated mantle temperatures expected beneath hotspots. The coesite-stishovite phase transition occurs in eclogite and/or carbonated silicate melt may form under certain conditions. For either mechanism to provide seismically observable impedance contrasts, the upper mantle must be locally enriched in basalt. Assessing the X-discontinuity global distribution reveals a correlation with the Large Low Velocity Provinces, suggesting that mantle plumes may entrain recycled basalt from the lowermost mantle.