A New Adjoint-state Full Waveform Tsunami Source Imaging Method

海啸是由海洋中大量水的位移产生的海洋重力波，通常是由复杂的地球物理过程造成的，包括破坏海洋地壳的浅层地震和海底滑坡。海啸会给沿海地区造成大量人员伤亡和经济损失，而这些地区的人口占世界人口的三分之一以上。因此，海啸预警是减少和减轻海啸危害的关键系统。为了更快速、准确地进行海啸预警，更好地了解海啸震源的复杂性，提出采用伴随状态法求解海啸震源的海底初始变形。该方法提高了震源过程的成像分辨率，产生的伪影较少，计算成本相对较小。本项目充分利用越来越多的海啸仪器，包括海底压力传感器和海岸潮汐计，并支持一名女学生的博士研究。研究结果通过会议和期刊进行分享，并扩展到公立学校和本科班级。传统的海啸震源反演要么基于有限断层滑动模拟，要么基于时间反演成像。这种反演方法受到断层参数或地壳刚度的不确定性的影响。此外，计算格林函数的计算量大，导致空间分辨率有限，阻碍了传统方法在海啸预警中的实时性。本文将目前最先进的伴随状态全波形反演方法从勘探地震学移植到海啸震源成像中。伴随状态法以较少的计算成本解决了初始水位-高程模式，有可能提高海啸预警速度，减少盲区。该方法不依赖于预定义的断层参数，适用于断层几何形状未知的海啸地震。该方法还能有效地处理密集的网格网格，并能解决海底滑坡或展断层上的次生破裂等小尺度海啸源。我们的研究结果将推动海啸科学和震源动力学的发展，并为改善海啸预警等实时应用奠定基础。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Tsunamis are oceanic gravity waves generated by the displacement of a large volume of water in the ocean, typically resulting from complex geophysical processes including shallow earthquakes that break the oceanic crust and seafloor landslides. Tsunamis can cause large number of casualties and economic loss in coastal regions, which are home to over a third of the world population. Tsunami warning is thus a crucial system to reduce and mitigate tsunami hazards. In order to perform a more rapid and accurate tsunami early warning as well as better understanding the complexity of tsunami source, the investigators propose to apply the adjoint-state method to solve for the initial seafloor deformation of the tsunami source. The method improves the imaging resolution of the earthquake source process and produces little artifacts with a relatively small computational cost. This project takes full advantage of increasing tsunami instrumentations including ocean-bottom pressure sensors and coastal tide gauges and supports the Ph. D work of a female student. The research results are being shared via conferences and journals, as well as outreach to public schools and in undergraduate classes.Traditional source inversion using tsunamis waves is based on either the finite-fault slip modeling or the time-reversal imaging. Such inversion methods suffer from the uncertainty of fault parameters or crustal rigidity. Moreover, the heavy computational burden of calculating Green's functions result in limited spatial resolution and hinders the real-time applicability of the traditional methods to tsunami early warning. In this work, we transplant the state-of-art adjoint-state full-waveform inversion method from exploration seismology to tsunami source imaging. The adjoint-state method solves the initial-water-elevation pattern with less computational cost, which potentially can improve the speed of tsunami early warning and reduces the blind zone. Our new method does not rely on pre-defined fault parameters and is suitable for tsunami-generating earthquakes with unknown fault geometry. This method also efficiently handles dense mesh grid and is capable of resolving small-scale secondary tsunami sources, such as the seafloor landslide or secondary ruptures on splay faults. Our results will advance tsunami science and earthquake source dynamics and set the stage to improve real-time applications such as tsunami early warning.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.