We present a general concept for evolutionary, collaborative, multiscale inversion of geophysical data, specifically applied to the construction of a first‐generation Collaborative Seismic Earth Model. This is intended to address the limited resources of individual researchers and the often limited use of previously accumulated knowledge. Model evolution rests on a Bayesian updating scheme, simplified into a deterministic method that honors today's computational restrictions. The scheme is able to harness distributed human and computing power. It furthermore handles conflicting updates, as well as variable parameterizations of different model refinements or different inversion techniques. The first‐generation Collaborative Seismic Earth Model comprises 12 refinements from full seismic waveform inversion, ranging from regional crustal‐ to continental‐scale models. A global full‐waveform inversion ensures that regional refinements translate into whole‐Earth structure.

We present a probabilistic seismic point source inversion, taking into account 3‐D heterogeneous Earth structure. Our method rests on (1) reciprocity and numerical wavefield simulations in complex media and (2) Hamiltonian Monte Carlo sampling that requires only a small amount of test models to provide reliable uncertainty information on the timing, location, and mechanism of the source. Using spectral element simulations of 3‐D, viscoelastic, anisotropic wave propagation, we precompute receiver side strain tensors in time and space. This enables the fast computation of synthetic seismograms for any hypothetical source within the volume of interest, and thus a Bayesian solution of the inverse problem. To improve efficiency, we developed a variant of Hamiltonian Monte Carlo sampling. Taking advantage of easily computable derivatives, numerical examples indicate that Hamiltonian Monte Carlo can converge to the posterior probability density with orders of magnitude less samples than the derivative‐free Metropolis‐Hastings algorithm, which we use for benchmarking. Exact numbers depend on observational errors and the quality of the prior. We apply our method to the Japanese Islands region where we previously constrained 3‐D structure of the crust and upper mantle using full‐waveform inversion with a minimum period of 15 s.

We present a full‐waveform tomographic model of the crust and upper mantle beneath the Japanese Islands region. This is based on the combination of GPU‐accelerated spectral‐element wavefield simulations, adjoint techniques, and nonlinear optimization. Our model explains complete seismic waveforms of events not used in the inversion in the period range from 20 to 80 s. Quantitative resolution analysis indicates that resolution lengths within the well‐covered areas are around 150 km in the horizontal and around 30 km in the vertical directions. In addition to the high‐velocity signatures of known lithospheric slabs in the region, our model reveals a pronounced low‐velocity anomaly beneath the volcanic island of Ulleung in the Sea of Japan, reaching −19% around 100 km depth. The Ulleung anomaly originates at or above the Pacific slab, rises vertically upward to the base of the Philippine Sea slab at ∼200 km depth, circumvents it in NW direction, and then significantly strengthens in the uppermost mantle above the Philippine Sea slab. Among the numerous hypotheses for the generation of low‐velocity anomalies in subduction systems, those invoking instabilities before or when a slab enters the transition zone seem most likely. The age and fast subduction of the Pacific slab may facilitate the transport of fluids into the transition zone. This may promote the reduction in viscosity and the onset of convective upwelling, aided by ambient mantle flow, such as return flow within the mantle wedge.

It is assembled from a long-wavelength 3-D global model into which several regional-scale tomographies are embedded. We illustrate the CSEM workflow of successive updating with two examples from Japan and the Western Mediterranean, where we constrain smaller scale structure using full-waveform inversion. Furthermore, we demonstrate the ability of the CSEM to act as a vehicle for the combination of different tomographic techniques with a joint fullwaveform and traveltime ray tomography of Europe. This combination broadens the exploitable frequency range of the individual techniques, thereby improving resolution. We perform two iterations of a whole-Earth full-waveform inversion using a long-period reference data set from 225 globally recorded earthquakes. At this early stage of the CSEM development, the broad global updates mostly act to remove artefacts from the assembly of the initial CSEM. During the future evolution of the CSEM, the reference data set will be used to account for the influence of small-scale refinements on large-scale global structure. The CSEM as a computational framework is intended to help bridging the gap between local, regional and global tomography, and to contribute to the development of a global multiscale Earth model. While the current construction serves as a first proof of concept, future refinements and additions will require community involvement, which is welcome at this stage already.

No abstract