干旱区生态环境知识资源中心

Postseismic displacements following great subduction earthquakes show significant long-wavelength and time-dependent patterns caused primarily by transient viscoelastic relaxation processes occurring broadly at depth. However, the Earth’s viscosity structure and time-dependent variations are still poorly understood, especially in the years immediately following a great earthquake. Here we investigate the spatiotemporal variation of mantle viscosity proximal and distal to the southern Andes using 8 years of continuous high-resolution GPS observations following the 2010 M-w 8.8 Maule earthquake in south Central Chile. We remove the potential influences of relocking and afterslip on far-field GPS displacements and estimate viscosities that can explain the 3-D displacements. The optimal viscosity structure exhibits a low-viscosity (similar to 10(18)Pa s) mantle beneath the Andean volcanic arc and high-viscosity (>10(22)Pa s) cratonic mantle, indicating a dependence of transient viscosity on temperature. Comparisons of the viscosity distributions at different times show that mantle viscosities increase with time throughout the study region. Viscosity increase is generally fastest in the mantle wedge beneath the Andes and slows down with increasing distance from the source region of the Maule earthquake. Such temporal viscosity evolution may indicate a stress dependence of the viscosity proximal to the rupture zone, while regions east of the Andes act as a relatively rigid body (i.e., cratonic mantle) with much higher viscosity. Our results thus suggest that both temperature structure and stress state contribute to spatiotemporal variations of the mantle viscosity. Heterogeneous spatiotemporal variations of viscosity seem to control the expansion and duration of the postseismic deformation and therefore the postseismic stress evolution.

Plain Language Summary We used continuous GPS positioning data to study the viscosity distribution at depth and its time variation following the 2010 M-w 8.8 Maule, Chile, earthquake. We find that the viscosities in the volcanic/mountain range regions, where deep rocks are expected hot, are relatively low (i.e., the materials are in weak state prone to flow easily), while the viscosities in far regions, where an ancient cold crust is expected, are relatively high (i.e., the materials are too strong to flow). This finding can, therefore, be related to the temperature distribution at depth. Furthermore, we find that the viscosities in all regions except the region of ancient crust increase with time and the rate of viscosity recovery decreases with the distance to the earthquake location. This phenomenon can, therefore, be related to the earthquake-induced stresses. Another interesting thing found in this study is that the ancient crust is likely not influenced much by the earthquake and behaves very strong in all the observation time. In all, we propose that both temperature and earthquake stresses control the rock strengths following a great earthquake. This study thus contributes to our understanding of the stress evolution at depth through the earthquake cycle and hence mechanisms of the earthquake happening.