Laura Ermert

Ambient seismic source inversion in a heterogeneous Earth - Theory and application to the Earth's hum†

The sources of ambient seismic noise are extensively studied both to better understand their influence on ambient noise tomography and related techniques, and to infer constraints on their excitation mechanisms. Here we develop a gradient-based inversion method to infer the space-dependent and time-varying source power spectral density of the Earths hum from cross correlations of continuous seismic data. The precomputation of wavefields using spectral elements allows us to account for both finite-frequency sensitivity and for three-dimensional Earth structure. Although similar methods have been proposed previously, they have not yet been applied to data to the best of our knowledge. We apply this method to image the seasonally varying sources of Earths hum during North and South Hemisphere winter. The resulting models suggest that hum sources are localized, persistent features that occur at Pacific coasts or shelves and in the North Atlantic during North Hemisphere winter, as well as South Pacific coasts and several distinct locations in the Southern Ocean in South Hemisphere winter. The contribution of pelagic sources from the central North Pacific cannot be constrained. Besides improving the accuracy of noise source locations through the incorporation of finite-frequency effects and 3-D Earth structure, this method may be used in future cross-correlation waveform inversion studies to provide initial source models and source model updates.

Seismic Noise Correlation on Heterogeneous Supercomputers

ABSTRACT We present a high‐performance tool for the computation of ambient seismic noise correlations on central processing unit (CPU) and graphic processing unit (GPU) clusters. This is intended to address emerging challenges in noise correlation studies with increasingly large data volumes. We propose a parallelization scheme and strategies to efficiently harness modern supercomputing resources, and we demonstrate that the use of GPUs can accelerate the computation of noise correlations by one order of magnitude or more compared with a homogeneous implementation on CPUs. In addition to reducing wall‐clock time, our tool enables on‐the‐fly computations of large noise correlation datasets, thereby eliminating the need for mass storage to archive results.

Towards Full Waveform Ambient Noise Inversion

We develop a method for the joint inversion of noise correlation functions for the distribution of noise sources and for Earth structure. The forward problem is free of assumptions required to equate noise correlations with Green functions and allows us to compute inter-station correlations for arbitrary distributions of noise sources in space and time. Using adjoint techniques, we design an iterative inversion scheme for noise sources and Earth structure based on waveform and energy differences as misfit functional. Starting from an initial model from a wave equation traveltime inversion, we recover the target velocity model with high accuracy. A key prerequisite is a good inference of the noise source distribution.

MULTI-SCALE/MULTI-DATA INVERSION FOR ELASTIC EARTH STRUCTURE - A CONCEPT

Complex interactions of smalland large-scale processes are characteristic for the physics of the Earth, and their proper quantification is key to the integration of interdependent geophysical systems that are today mostly treated as isolated. Inferring Earth structure over a wide range of scales is the long-standing goal of seismic tomography. While much progress has been made in recent years, tomographic resolution remains limited by our inability to model and invert seismic wave propagation across the complete observable frequency band with the currently available computational resources. Here we propose a new concept for multi-scale seismic tomography intended to resolve Earth structure from local to global scales, including mantle as well as detailed crustal features. For this we develop a multi-scale full waveform inversion technique that assimilates complete teleseismic and regional seismograms in a broad frequency band. Being based on spectralelement modelling and adjoint techniques, our method simultaneously solves multiple regionaland continental-scale inverse problems in order to jointly resolve Earth structure with resolving lengths ranging from around 20 to more than 5000 km. To further increase the exploitable frequency band beyond what can be modelled numerically, we combine full waveform inversion with classical ray tomography that assimilates the arrival times of high-frequency body waves into the tomographic model. This combination results in an improved resolution of Earth structure, especially below 300 km depth. We apply our method to Europe and Western Asia, where resolution is particularly high beneath the North Atlantic, the Western Mediterranean and Anatolia. Quantitative resolution analysis based on second-order adjoints, as well as comparisons with observed ambient noise correlations, allow us to assess the quality and predictive power of the final model.