We present a novel geodynamic approach that can potentially tighten existing constraints on mantle rheology. This new approach, which we call probabilistic geodynamic modeling, is applied here to the rheology of the upper mantle. We combine the numerical modeling of plate‐driven corner flow and the seismic observation of radial anisotropy, aiming to reduce substantial uncertainties associated with experimentally derived flow laws, but our results also highlight the complex competition among different deformation mechanisms under mantle conditions. Despite the remaining rheological uncertainty, our study suggests that significant background shear flow is required near the lithosphere‐asthenosphere boundary to explain the strong radial anisotropy observed at 100–200 km depth underneath the Pacific plate, and the plausible nature of this background flow is characterized using our new probabilistic approach. Our analysis also provides a new insight into the asthenospheric water content and the grain size distribution in the upper mantle, but these results are also subject to nontrivial nonuniqueness. The merit of our probabilistic approach lies in its ability to assess the extent of such nonuniqueness, and we demonstrate this by quantifying the robustness of some of our results.

中文翻译：

---

实验岩石力学，地震学和地球动力学的协同作用揭示了仍然难以捉摸的上地幔流变学

我们提出了一种新颖的地球动力学方法，可以潜在地加强对地幔流变学的现有限制。我们将这种称为概率地球动力学建模的新方法应用于上地幔的流变学。我们将板驱动角流的数值模型与径向各向异性的地震观测相结合，旨在减少与实验得出的流定律相关的大量不确定性，但我们的结果也强调了地幔条件下不同变形机制之间的复杂竞争。尽管仍存在流变学上的不确定性，我们的研究表明，岩石圈-软流圈边界附近仍需要大量的背景切变流，以解释在太平洋板块下100-200 km深度处观测到的强烈的径向各向异性，使用我们的新概率方法来表征背景流的合理性质。我们的分析还为软流圈水含量和上地幔的粒度分布提供了新的见解，但这些结果也受到非平凡的非唯一性的影响。我们的概率方法的优点在于它能够评估这种非唯一性的程度，我们通过量化某些结果的鲁棒性来证明这一点。