Testing the Thermal Shear Instability Hypothesis for Deep Slab Seismicity

深地震发生在地球表面以下100至680公里处的寒冷构造板块内，这些板块正在下沉到地球内部。对于这些地震的原因，有一种流行的理论，称为转换断层，但这种理论无法解释所有与深源地震有关的观测结果。最近，研究表明，另一种机制，称为热剪切不稳定性（TSI），或两种机制的组合，可能更好地符合观测结果。这项研究将使用下沉构造板块（“板”）的模拟，以确定温度，应力和变形率在不同的深度/位置的板。然后，这些条件将被用作可以对TSI进行建模的第二类模拟的起始条件。我们预计，对于某些条件TSI（即，“地震”）将发生，而在其它条件下则不会发生。因此，使用第二类模型的输出，我们可以绘制出在板TSI是一个可能的深源地震的机制。如果我们的假设是正确的，那么对引起深源地震的机制的理解的转变（从一种机制到多种潜在机制）可能会导致新的研究，旨在将地震观测与深源地震的破裂特性直接联系起来。从广义上讲，这些结果将进一步加深我们对导致地震破裂的过程和条件的理解。最近对热剪切不稳定性（TSI）所需条件的模拟和实验室测量表明，TSI可能是俯冲构造板块中深度达150 km左右的深地震的可行机制。同时，对150-680 km深震震级-频度分布的分析也被用来论证TSI在触发深震，特别是在较暖的板块中的作用。该项目将测试TSI作为在100-680公里深处的俯冲构造板块内触发深地震的机制的可行性。这将分三步进行。首先，我们将运行多个剖面和俯冲带的二维粘弹塑性模型，这些剖面和俯冲带具有不同的几何形状、板块年龄、俯冲速率和深地震活动的速率/空间变异性。该模型将使用粘弹塑性流变学，并将运行0.1-1.0 my，以确定准稳态空间分布和弹性应力的大小，以及板中的总应变率。其次，我们将使用来自2D板坯模型的压力、温度、应力和应变率条件的范围单独运行1D TSI模型，以确定在板坯中存在的条件下发生TSI。TSI模型将使用与二维俯冲模型相同的流变学。这一比较将说明板状TSI是深源地震的潜在触发机制。最后，我们将比较特定位置的地震观测（空间分布，震源机制，震级，b值）的模型结果（空间分布的TSI，断层方向，震级的估计和地震活动性统计的几何约束）。这种组合模型的结果与观测结果的比较将证明我们的模拟如何捕捉到整体变形的板在地震破裂的短时间尺度通过较长的时间尺度，确定目前的应力状态在slab.This奖项反映了NSF的法定使命，并已被认为是值得支持的，通过评估使用基金会的智力价值和更广泛的影响审查标准。

EDITDeep earthquakes occur 100 to 680 km below Earth’s surface within cold tectonic plates that are sinking back into the Earth’s interior. There has been one prevailing theory for the cause of these earthquakes, called transformational faulting, but this theory is not able to explain all the observations related to deep earthquakes. Recently, research has indicated that another mechanism, called thermal shear instability (TSI), or a combination of both mechanisms, may be better able to match the observations. This research will use simulations of sinking tectonic plates (“slabs”) to determine the temperature, stresses, and rates of deformation at differentdepths/locations in the slab. These conditions will then be used as the starting conditions of a second type of simulation that can model TSI. We expect that for some conditions TSI (i.e., an “earthquake”) will occur, while at other conditions it will not. Therefore, using the output of the second type of model we can map out where in the slab TSI is a possible mechanism for deep earthquakes. If our hypothesis is correct, the shift in understanding of the mechanism causing f deep earthquakes (from one, to multiple potential mechanisms) would likely lead to new research aimed at directly linking seismic observations to the rupture properties of deep earthquakes. Broadly speaking, these results will further our understanding of the processes and conditions that lead to earthquake rupture.Recent modeling of, and laboratory measurements on, the conditions needed for thermal shear instability (TSI) have demonstrated that TSI may be a viable mechanism for deep earthquakes in subducting tectonic plates at depths up to around 150 km. At the same time, analysis of the magnitude-frequency distribution of deep earthquakes from 150-680 km has also been used to argue that TSI plays a role in triggering deep earthquakes, especially in warmer slabs. This project will test the viability of TSI as a mechanism for triggering deep earthquakes within subducting tectonic plates at depths of 100-680 km. This will be done in a three-step process. First, we will run 2D visco-elasto-plastic models for multiple profiles and subduction zones with different geometry, plate ages, rates of subduction, and rates/spatial variability of deep seismicity. The models will use a visco-elasto-plastic rheology and will be run for 0.1-1.0 my to determine a quasi-steady state spatial distribution and magnitude of elastic stresses, and total strain rate in the slab. Second, we will separately run 1D TSI models using the range of pressure, temperature, stress, and strain-rate conditions from the 2D slab models to determine at which conditions, present in the slab, TSI occurs. The TSI models will use the same rheology as the 2D subduction models. This comparison will demonstrate where in the slab TSI is potential triggering mechanism for deep earthquakes. Finally, we will compare location-specific earthquake observations (spatial distribution, focal mechanisms, magnitudes, b-values) to the model results (spatial distribution of TSI, fault orientations, estimates of magnitudes and geometric constraints on seismicity statistics). This comparison of the combined model results to observations will demonstrate how well our simulations capture the overall deformation of the slab at the short timescales of earthquake rupture up through the longer time-scales that determine the present-day stress-state in the slab.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.