本项目拟以地震学为主要手段来探测深俯冲带附近的地球物理性质，辅以相关矿物在相应温-压条件下的实验室数据，而聚焦于地球深部水循环的重大问题。研究重点在上下地幔转换带与深俯冲板块的相互作用。这里是地幔主要矿物产生固态相变化的范围；在矿物含水、脱水机理，流变性质，深源地震的触发，及地球动力学方面意义特别重大。拟赴全球最重要的俯冲带采集崭新数据, 扩大国内已有的基础，以争取突破性进展。工作内容拟集中在全球俯冲速率最快，深源地震占全球三分之二，俯冲海洋板块形貌最奇特的汤加-克马得经典地区，以超宽频带的地震仪数组采集高分辨率、高密度的详实数据。再以先进的方法分析地震波速度的三维分布，剪波各向异性的变化，深源地震三维分布与地震节面解的规律等可以鉴别名义无水矿物是否存在的参数。最后与我国东北地区已有的结果对比，以期可靠的制约快速俯冲板块在不同深度含水、脱水的程度及机理，从而进一步了解地球深部水循环的谜团。

We propose to collect high-resolution, dense-spaced, broadband seismic array data near the Tonga-Kermadec trench in order to provide additional constraints on the process of deep subduction of the oceanic lithosphere and its interaction with the mantle transition zone. Through joint interpretations with laboratory data on the physical properties of mantle minerals at temperatures and pressures appropriate for the transition zone of the mantle, including those of nominally anhydrous minerals, and comparisons with prior results from the deep-subducted Pacific plate beneath northeastern China, our goals are twofold. First is to resolve whether deep subduction of the oceanic lithosphere is an important pathway for transporting water into the deep mantle in particular, and second, to advance our understanding of deep subduction processes and the Earth's deep water cycle in general. The Tonga-Kermadec deep subduction zone holds the key to these important issues because old (>100 Ma, and therefore cold) oceanic lithosphere is being subdcuted at an astonishing rate of over 200 mm/yr；and a full two-thirds of all deeps earthquakes in the world occur along this subduction zone. Furthermore, the abundant deep-seismicity outlines an intriguing, highly contorted configuration of subducted lithosphere, including the so-called outboard earthquakes that occur over 1,000 km away from the typical Wadati-Benioff zone, which, in turn, indicate "ghost" slab remnants that are anisotropic and seismogenic but not associated with elevated seismic wave speeds. .Using advanced techniques of analyzing broadband seismic array data, such as modeling of potentially anisotropic, triplicate waveforms; and travel-time tomography that includes effects of wave front healing (finite-frequency kernels) and a data-adaptive, multi-scale treatment of the model space, we anticipate high-resolution results on three-dimensional distribution of P- and S-wave-speed anomalies, variations of seismic anisotropy with depth, and patterns of focal-mechanisms and deep seismicity. Together with mineral physics data and tectonic history of subduction of the western Pacific, a holistic interpretation of the seismic constraints is expected to yield new insights on how subducted lithosphere interacts with the deep mantle, including the exchange of water at depth.