Based on SEMUCB‐WM1 tomographic model, validated by other recent models, and fluid mechanics constraints, we show that the large low shear velocity provinces (LLSVPs) present at the base of the Earth's mantle beneath the Pacific and Africa do not extend as compact, uniform structures very high above the core‐mantle boundary. In contrast, they contain a number of well‐separated, low‐velocity conduits that extend vertically throughout most of the lower mantle. The conceptual model of compact piles, continuously covering the areal extent of the LLSVPs, is therefore not correct. Instead, each LLSVP is composed of a bundle of thermochemical upwellings probably enriched in denser than average material. It is only when the tomographic model is filtered to long wavelengths that the two bundles of plumes appear as uniform provinces. Furthermore, the overall shape of the LLSVPs is probably controlled by the distribution of subducted slabs, and due to their thermochemical nature, the position of both LLSVPs and individual upwelling dynamics should be time dependent. There is also evidence for smaller plumes originating near the CMB in the faster than average regions of the voting map of Lekic et al. (2012, https://doi.org/10.1016/j.epsl.2012.09.014) as well as other, barely resolved, weaker plumes within the LLSVPs. These finer‐scale features are starting to be resolved tomographically owing to improvements in full waveform modeling of body waves, including diffracted S waves (Sdiff) and waves multiply reflected on the core‐mantle boundary (ScS) and their codas.

中文翻译：

---

缩小LLSVP：地幔热化学羽成束，而不是停滞不前的“堆”

基于SEMUCB-WM1层析成像模型（已通过其他最新模型验证）以及流体力学约束，我们表明，位于太平洋和非洲下方的地幔底部存在的大型低剪切速度省（LLSVP）并不像紧凑的那样延伸，高于地幔边界的均匀结构。相反，它们包含许多分离良好的低速导管，这些导管在整个下地幔的大部分垂直延伸。因此，连续覆盖LLSVP面积的紧凑桩的概念模型是不正确的。取而代之的是，每个LLSVP都由一束热化学上升流组成，这些热化学上升流的密度可能高于平均水平。只有当层析成像模型被过滤到长波长时，两束羽才会显示为均匀的省份。此外，LLSVP的整体形状可能受俯冲板的分布控制，并且由于其热化学性质，LLSVP的位置和上升流动力学都应与时间有关。也有证据表明，莱比奇（Lekic）等人的投票图的平均区域中，较慢的羽流起源于中巴附近。（2012，https://doi.org/10.1016/j.epsl.2012.09.014）以及LLSVP内其他几乎未解决的较弱羽状流。由于体波的完整波形建模（包括衍射）的改进，这些更精细的特征已开始在层析成像上得到解决 也有证据表明，较小的羽流起源于CMB附近，其速度比Lekic等人投票地图的平均区域快。（2012，https://doi.org/10.1016/j.epsl.2012.09.014）以及LLSVP中其他几乎未解决的较弱羽状流。由于体波的完整波形建模（包括衍射）的改进，这些更精细的特征已开始在层析成像上得到解决 也有证据表明，莱比奇（Lekic）等人的投票图的平均区域中，较慢的羽流起源于中巴附近。（2012，https://doi.org/10.1016/j.epsl.2012.09.014）以及LLSVP内其他几乎未解决的较弱羽状流。由于体波的完整波形建模（包括衍射）的改进，这些更精细的特征已开始在层析成像上得到解决S波（Sdiff）和波在芯幔边界（ScS）及其尾气上反射。