这篇综述讨论大空间、大时间尺度的地幔动力学近几十年的发展和现状，着重讨论了相关的观测及其动力学意义.这些观测包括现在地球的板块运动的基本特性，中、长波重力异常及大地水准面异常，地震层析成像得到的地幔结构，以及过去10亿年超级大陆Pangea和Rodinia的形成、裂解和演化，及火山岩浆活动.关于地球动力学模型的讨论是围绕着这些相关的观测而进行的.涉及到的一些主要问题包括以下.第一，地幔动力学研究显示，地震层析成像得到的下地幔的二阶结构（比如核幔边界附近的LLSVP结构），和俯冲带的快速异常体，可以解释为过去1亿年左右的板块运动和地幔对流的结果；第二，地幔三维结构作为地幔对流的驱动力，是导致中、长波重力及大地水准面异常的直接原因；结合地幔动力学模拟，观测的大地水准面异常对地幔黏性结构提供了强有力的约束，很可靠的结果之一是下地幔的黏性比上地幔要高至少一个量级，并且最近的研究确定软流圈的存在；第三，过去10亿年大陆块体经历过的Rodinia和Pangea两期超级大陆的形成和破裂是地幔动力学在地表的反映.地幔结构在Pangea形成过程中是一阶结构（即一个半球是冷的下降流，而另一个半球是热的上涌流）主导的，而现在的二阶为主导的地幔结构是Pangea形成后，破裂前或破裂过程中才形成的；地幔动力学和其他研究支持地幔结构在一阶和二阶间转换的1-2-1模型；第四，板块构造在地球上的起源和动力机制依然是充满争议和不确定的课题，但是这些问题同时也是重要的地球动力学基本问题.

地幔动力学 Abstract:

This article provides a review on the studies of large temporal and spatial scale dynamics of the Earth's mantle. The review focuses on relevant observations and their geodynamic interpretations and implications. These observations include present-day Earth's plate tectonics, long- and intermediate-wavelength geoid and gravity anomalies, and mantle seismic structures, as well as important tectonism and magmatism that have happened in the last one billion years, associated with the formation and breakup of supercontinents Pangea and Rodinia. Much of the discussion is centered on how these observations have motivated geodynamic studies and modeling that seek to understand and interpret the observations. This review covers four topics. The first is on the primary characteristics of mantle seismic structure and their dynamic origin. The present-day Earth's mantle is predominated by long-wavelength structures (i.e., degree-2 in the lower mantle and LLSVPs near the core-mantle boundary) and linear structures in subduction zones, both of which can be interpreted as a result of mantle convection modulated by surface plate motion history in the last 100 million years. The second is on the long- and intermediate-wavelength geoid and gravity anomalies and their dynamic interpretation. The geoid anomalies are explained by mantle flow that is driven by buoyancy associated with the mantle structure. Such studies indicate that the upper mantle is at least one magnitude weaker than the lower mantle and strongly suggest the existence of a weak asthenosphere. Third, the cyclic process of formation and breakup of supercontinents Pangea and Rodinia is surface manifestation of time-dependent mantle convection. During supercontinent formation and its early stage, mantle structure is predominately degree-1 with cold downwellings in one hemisphere and hot upwellings in the other hemisphere. However, the degree-1 structure starts to transition to degree-2 mantle structure with two major antipodal upwelling systems (e.g., the present-day Earth) in the late stage of a supercontinent, leading to supercontinent breakup. Abundant observational and dynamic evidence support the 1-2-1 model for supercontinent cycle and mantle structure evolution. The fourth is on the origin of plate tectonics and long-term thermal evolution of the Earth which is a fundamentally important but also controversial topic in the studies of earth science.

Key words: Mantle convection Mantle dynamics Supercontinents Plate tectonics Evolution of the Earth Figure 1.

S-wave speed anomalies of the Earth′s mantle from seismic tomography (Ritsema et al., 2001) at depths of 1800 km (a) and 2800 km (b), and their normalized depth-dependent power spectra (c). Note that the slow wave speed anomalies under Africa and Pacific (e.g., in Fig. 1b ) represent LLSVP. The power spectra show degree-2 dominant mantle structure throughout the mantle

Figure 4.

Distributions of hotspot volcanism, LIP, and S-wave speed anomalies ( Becker and Boschi, 2002 ) (a), temporal variations of magmatism for the last 3 billion years ( Ernst and Bleeker, 2010 ) (b), and temporal dependence of LIP for the last 500 million years ( Torsvik et al., 2008a ) (c). Fig. 4b also shows major tectonic events, and Fig. 4c marks latitudes of LIP′s eruption sites and possible connection of Africa and Pacific LLSVPs with Pangea

Figure 7.

3-D representation of S-wave speed anomalies of the mantle ( Ritsema et al., 2011 ) viewed from Pacific (a) and African (b) hemispheres, and corresponding thermochemical structures (c, d) from mantle convection models using imposed plate motion history for the last 120 million years. Modified from McNamara and Zhong (2005a)

England P. 2018. On shear stresses, temperatures, and the maximum magnitudes of earthquakes at convergent plate boundaries. J. Geophys. Res. : Solid Earth , 123(8): 7165-7202. http://onlinelibrary.wiley.com/doi/10.1029/2018JB015907 Grand S P, van der Hilst R D, Widiyantoro S. 1997. Global seismic tomography: A snapshot of convection in the Earth. GSA Today , 7(4): 1-7. http://geosociety.org/gsatoday/archive/7/4/pdf/i1052-5173-7-4-1.pdf Hager B H, Richards M A. 1989. Long-wavelength variations in Earth's geoid: physical models and dynamical implications. Philos. Trans. Roy. Soc. Lond. Ser. A: Math. Phys. Sci. , 328(1599): 309-327. http://sas2.elte.hu/tg/msc_gravi/NEW_Hager_Richards_1989_PhilTransRSL.pdf Masters G, Laske G, Bolton H, et al. 2000. The Relative Behavior of Shear Velocity, Bulk Sound Speed, and Compressional Velocity in the Mantle: Implications for Chemical and Thermal Structure. //Karato S I, Forte A, Liebermann R, et al eds. Earth's Deep Interior: Mineral Physics and Tomography from the Atomic to the Global Scale. Washington, DC: American Geophysical Union, 63-87. Figure 1.

S-wave speed anomalies of the Earth′s mantle from seismic tomography (Ritsema et al., 2001) at depths of 1800 km (a) and 2800 km (b), and their normalized depth-dependent power spectra (c). Note that the slow wave speed anomalies under Africa and Pacific (e.g., in Fig. 1b ) represent LLSVP. The power spectra show degree-2 dominant mantle structure throughout the mantle

Figure 2.

S-wave speed anomalies for different subduction zone depth cross sections

Figure 3.

The observed geoid anomalies of degrees 2 to 12 (a), and degrees 4 to 12 (b), and the corresponding geoid anomalies from mantle convection model (c, d) ( Mao and Zhong, 2021a )

Figure 4.

Distributions of hotspot volcanism, LIP, and S-wave speed anomalies ( Becker and Boschi, 2002 ) (a), temporal variations of magmatism for the last 3 billion years ( Ernst and Bleeker, 2010 ) (b), and temporal dependence of LIP for the last 500 million years ( Torsvik et al., 2008a ) (c). Fig. 4b also shows major tectonic events, and Fig. 4c marks latitudes of LIP′s eruption sites and possible connection of Africa and Pacific LLSVPs with Pangea

Figure 5.

A schematic drawing of different mantle convection models

Figure 6.

Present-day plate configuration and plate motions (i.e., arrows), and a viscosity model at 80 km depth where the light blue color represents weak plate boundaries ( Mao and Zhong, 2021a )

Figure 7.

3-D representation of S-wave speed anomalies of the mantle ( Ritsema et al., 2011 ) viewed from Pacific (a) and African (b) hemispheres, and corresponding thermochemical structures (c, d) from mantle convection models using imposed plate motion history for the last 120 million years. Modified from McNamara and Zhong (2005a)

Figure 8.

Supercontinents Pangea at 195 million years ago (a) and Rodinia at 750 million years ago (b). Modified from Zhong et al. (2007)

Figure 9.

Thermal structure at 2750 km depth from mantle convection model with plate motion history for the last 580 million years at 330 million years ago (i.e., when Pangea is assembled) (a), 195 million years ago (b), and the present-day (c). Modified from Zhang et al. (2010)

Figure 10.

3-D thermal structure from different mantle convection models (blue and yellow isosurfaces represent cold downwellings and hot upwellings, respectively)

Figure 11.

Dominant convective wavelengths versus averaged lithospheric viscosity from 3-D mantle convection models with dynamic boundary conditions (i.e., free-slip)

Figure 12.

A model for plate tectonics generation based on damaging theory ( Bercovici and Ricard, 2014 )