Koelemeijer, Paula JSchuberth, Bernhard S ADavies, D. RhodriDeuss, ArwenRitsema, Jeroen2019-03-250012-821Xhttp://hdl.handle.net/1885/157233Lower mantle tomography models consistently feature an increase in the ratio of shear-wave velocity (VS) to compressional-wave velocity (VP) variations and a negative correlation between shear-wave and bulk-sound velocity (VC) variations. These seismic characteristics, also observed in the recent SP12RTS model, have been interpreted to be indicative of large-scale chemical variations. Other explanations, such as the lower mantle post-perovskite (pPv) phase, which would not require chemical heterogeneity, have been explored less. Constraining the origin of these seismic features is important, as geodynamic simulations predict a fundamentally different style of mantle convection under both scenarios. Here, we investigate to what extent the presence of pPv explains the observed high VS/VP ratios and negative VS–VC correlation globally. We compare the statistical properties of SP12RTS with the statistics of synthetic tomography models, derived from both thermal and thermochemical models of 3-D global mantle convection. We convert the temperature fields of these models into seismic velocity structures using mineral physics lookup tables with and without pPv. We account for the limited tomographic resolution of SP12RTS using its resolution operator for both VS and VP structures. This allows for direct comparisons of the resulting velocity ratios and correlations. Although the tomographic filtering significantly affects the synthetic tomography images, we demonstrate that the effect of pPv remains evident in the ratios and correlations of seismic velocities. We find that lateral variations in the presence of pPv have a dominant influence on the VS/VP ratio and VS–VC correlation, which are thus unsuitable measures to constrain the presence of large-scale chemical variations in the lowermost mantle. To explain the decrease in the VS/VP ratio of SP12RTS close to the CMB, our results favour a pPv-bearing CMB region, which has implications for the stability field of pPv in the Earth's mantle.PK is funded by a Junior Research Fellowship from University College, Oxford and acknowledges support from an ETH Zürich Postdoctoral Fellowship (ETH/COFUND FEL-25 13-2, grant agreement number 0-20997-14). DRD is funded by an ARC Future Fellowship (FT140101262) and JR is supported by NSF grant EAR-1644829. AD was also funded by the ERC under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 681353 ATUNE) and a Vici award number 016.160.310/526 from the Netherlands Organisation for Scientific Research (NWO). Further funding was provided by Pembroke College, Cambridge and the European Research Council (ERC) under the European Community’s Seventh Framework Programme (FP7/2007-2013)/ERC grant agreement number 204995. Mantle dynamics simulations were undertaken on the NCI National Facility in Canberra, Australia, which is supported by the Australian Commonwealth Government.13 pagesapplication/pdfen-AU© 2018 Elsevier B.Vseismic tomographytomographic filteringcomposition of the mantlegeodynamic modellingmineral physics201810.1016/j.epsl.2018.04.0562019-03-12