The uppermost layer of Earth's lithosphere is magnetized. Rock magnetization is a valuable archive as it carries information about rocks' properties, Earth's tectonic history and evolution of its core magnetic field. Laboratory measurements of rock samples enable us to extract some of this information. Collecting rocks, however, allows only a very sparse and inhomogeneous coverage of the lithosphere. An alternative way to study Earth's magnetization is through the magnetic signal it generates, known as the lithospheric magnetic field. Although there exist global lithospheric magnetic field models of high spatial resolution, this path has not yet been very fruitful because of an important limitation: only part of the magnetization is visible, i.e., it produces an observable magnetic field signal. We refer to the remaining part of the magnetization as the hidden magnetization, in the sense that it produces no observable magnetic field. Here, we show that most of the hidden magnetization can be uniquely recovered from a global lithospheric magnetic field model under the assumption of purely induced magnetization. For the spatial scales we are dealing with, which go down to 200 km, this is a realistic assumption, particularly over the continents. To recover the remaining part of the hidden magnetization we rely on a priori information and in particular on a given vertically integrated susceptibility model (Hemant&Maus, 2005).

At the core of our method lies a newly derived formula that connects the magnetic susceptibility and the inducing magnetic field to the visible and hidden parts of the magnetization (Vervelidou&Lesur, 2018). In particular, our methodology consists in converting a lithospheric magnetic field model to a visible magnetization model according to the methodology of Vervelidou et al., 2017, then inferring the underlying susceptibility from the visible magnetization model by solving a regularized inverse problem and finally estimating the hidden magnetization through a forward calculation.

Our input lithospheric magnetic field model is based on satellite, airborne and marine magnetic field measurements. In particular, for spatial scales down to 450 km, we used the Dedicated Lithospheric Field Model (Thébault et al., 2016), a product of the Swarm multi-satellite mission. For shorter spatial scales, down to 200 km, we used the Spherical Harmonic model obtained from the second version of the World Digital Magnetic Anomaly Map (WDMAM) (Lesur et al., 2016).

Our results suggest that more than 60% of Earth’s magnetization is hidden. At middle and high latitudes, where the main source of information is the input lithospheric magnetic field model, Earth’s hidden magnetization is very similar to the visible magnetization, both in intensity and in shape. At low latitudes, the magnetization relies heavily on the a priori susceptibility model used in the inversion scheme. There, our results show that several structures of the a priori susceptibility model give rise only to hidden magnetization. Recovering the entire magnetization allows new magnetized structures to become apparent and other, known ones, to become more accurately delineated. In both cases a more accurate link to the underlying Earth processes becomes possible.

The C-response is a conventional transfer function (TF) in global electromagnetic induction studies which is determined from observations of local variations in the vertical and horizontal components of the Earth's geomagnetic field. The C-response concept is based on the assumption that the source of the variations (large-scale magnetospheric ring current) is well described by a single, first zonal, spherical harmonic. However, there is growing evidence for a more complex structure of the magnetospheric source, and unaccounted-for complexity of the source may introduce uncontrolled errors in the estimated C-responses. To overcome the problem, Püthe et al. (2015) introduced arrays of alternative, global-to-local TFs that can handle a complex spatial structure of the magnetospheric source. These TFs relate local variations in vertical magnetic field with spherical harmonic coefficients describing the source. In this work, we estimate and invert new TFs estimated at a number of worldwide distributed geomagnetic observatories to map lateral variations of electrical conductivity in the Earth's mid mantle (~ 400-1500 km depth). The obtained conductivity profiles are subsequently interpreted in terms of mantle thermo-chemical structure based on self-consistent thermodynamic calculation of mineral phase equilibria, electrical conductivity databases, and probabilistic inverse methods.

Reference:
Christoph Püthe, Alexey Kuvshinov, Nils Olsen; Handling complex source structures in global EM induction studies: from C-responses to new arrays of transfer functions, Geophysical Journal International, Volume 201, Issue 1, 1 April 2015, Pages 318–328, https://doi.org/10.1093/gji/ggv021

Surface observations of aseismic creep on the faults north of San Francisco Bay in California are sparse. Creep-affected infrastructure, e.g. fractured roads and offset kerbstones, give hints of its occurrence, yet the lateral and depth extents of creep on faults in the region remain poorly known. Better knowledge of the distribution of creep is helpful for identifying areas with greater or lesser potential for hazardous earthquake rupture (presence of creep being consistent with a reduced hazard, and vice-versa). Such knowledge will also help us place constraints on the underlying mechanism for creep, and in particular to identify the geological unit(s) that may cause it.

Characteristic repeating earthquake sequences, groups of identical microearthquakes whose waveforms have mutually high cross-correlation coefficents at multiple seismic stations, are one possible line of evidence consistent with creep. These sequences are thought to represent small asperities on the fault surface that are surrounded and periodically reloaded by creep. We detect ~200 such sequences, suggesting that creep should be pervasive on several of the region's active faults, notably the Maacama and northern Rodgers Creek faults, at depths of 4-9 km.

In order to constrain the behaviour of the shallow portions of these faults, we use persistent scatter InSAR (PSI) data. Preliminary efforts using archival data from the ERS and Envisat missions, have showed that PSI can highlight shallow deformation due to fault creep in areas where damaged infrastructure has been identified, but lacked the spatial resolution to map it more broadly, predominantly due to the irregular temporal sampling of observations and the issue of vegetation.

The Sentinel-1 mission provides a game-changing opportunity to improve our fault deformation mapping, given its excellent temporal and spatial coverage in California. Here we present results of a 130 acquisition dataset from Sentinel-1, spanning 2015 to 2018. We use the JPL ISCE software to align each to a common ‘super master’ date, and then produce over 400 interferograms with baselines < 100 m and timespans of less than one year. We then input these interferograms into the small baseline subset code implemented in the StaMPS/MTI software; this approach is demonstrated to be most effective in rural areas, such as rural northern California. We estimate relative velocity profiles at regular short intervals along the fault trace by making local double differences across the fault. As tropospheric noise cancels out over short distances (<10 km), we interpret these velocities in terms of shallow creep (a step change across the fault) or locking (a smooth velocity profile), and hence provide a spatial map of the creep distribution.

The gravity field anomaly over North America shows a strong negative gravity anomaly close to -40 mGal (using a global gravity field model up to spherical harmonic degree 60). The anomaly is believed to be the result of dynamic effects including incomplete Glacial Isostatic Adjustment (GIA) and dynamic topography due to past subduction of the Farallon slab. However, variations in crustal thickness and variations in mantle density also play a role. Several studies exist which have identified which of the components is the dominant cause for the negative gravity anomaly, but the full uncertainty range for all components has not been investigated. Here, we investigate the contribution of mantle density variations and GIA. The aim is to find out how large the remaining gravity anomaly is, and if it can provide a constraint on dynamic topography, which is the component with the largest uncertainty.

We use a temperature model (WINTERC3.2) derived from surface waveform data, surface elevation and surface heat flow. The model accounts for variation in composition in the lithosphere and sub-lithosphere (Al₂O₃, FeO). At 200 km depth below North America temperature varies by several hundreds of degrees, which translates into significant density anomalies. We use a GIA model, starting with a model with a radial viscosity profile. Gravity rate derived from GRACE time series and uplift rates from GPS measurements are used to constraint viscosity. The best fit model has a modest increase in viscosity from upper to lower mantle of a factor of 4 to 8. The resulting GIA induced topography varies from -70 to -140 m. This means that topography must be corrected by this amount before it can be used in isostatic calculations. The aim for future work is to make numerical models consistent with eachother by using the WINTERC temperature model as input for both GIA and dynamic topography models.

The satellite mission GOCE delivered gravity gradient data with a global homogeneous coverage. In previous studies, it has been shown that satellite gravity gradients are more sensitive to crustal structures than the conventional vertical gravity field. This allows crustal (and lithospheric) modelling in remote areas, which are sparsely covered by active seismology. The cratonic cores of South America (Amazonia and Sao Francisco) and West Africa, which have a common tectonic history prior to Gondwana breakup, are such areas with sparse terrestrial data coverage. The crust-mantle boundary is of main interest as it holds the main density contrast in the lithosphere, expressed by a jump in density due to the transition from crust to mantle rocks. Thus, the Moho depth should be resolvable using satellite gravity data.

We developed an inverse scheme, where gravity gradient components are inverted for the Moho depth. The inverse problem is solved with the classical Gauss-Newton algorithm. The Moho depth undulations are discretized into tesseroids. With this configuration, the full tensor gravity of the Moho model can be calculated using tesseroids software, which accounts for the spherical shape of the earth. The initial gravity gradients are corrected for the gravitational effect of topography and sediments. Furthermore, we account for far-field topographic and isostatic effects that have to be considered when working with global gravity gradients on a regional scale.

The inversion is focused on the region of the cratons of South America and West Africa. The final Moho model is a compromise between the gravity-inverted Moho depth and the existing seismic data in the study area. The depths are correlated with tectonic provinces and show for example that Sao Francisco has a significant higher crustal thickness than Amazonia. The internal architecture of Amazonia reflects deeper Moho depths for Archaean units.

Together with satellite residual mantle gravity anomalies and gradient-derived curvature products, the inverted Moho depths and gravity gradients are implemented in the plate reconstruction software GPlates. By rotating the input fields back in time, we are able to study the link between the different cratonic units within Gondwana assembly and prior stages.

Anna Maria Marotta, Arcangela Bollino, Federica Restelli, Gabriele Cambiotti, Roberto Sabadini and the GRAVSEIS Group*

Department of Earth's Sciences, Università degli Studi di Milano, c/o L. Cicognara 7, 20129 Milan, Italy

Tectonics is the source of mass anomalies within the Earth, leading to gravity anomalies in a wide spectrum of spatial and temporal scales.
Different amplitudes and time scales characterize the different tectonic processes, being each phenomenon categorized by its own time scale and wavelength, from short wavelength ones associated with shallow surface features, to long wavelength ones due to mantle convection, and from static and quasi-static ones from slow processes to the fast ones associated with earthquakes.
This contribution focuses on the results of a series of numerical models quantifying the gravitational effects of the stationary and slow time varying active Solid Earth processes responsible for the density structures that the Next Generation Gravity Mission (NGGM) is expected to sample, recognizing the static gravity field, as well as the slowly varying one, due to tectonics. The gravitational effects of tectonics are discussed in terms of their amplitude - in mGal for the gravity anomaly and in Eötvös for the gravity gradients – in terms of their amplitude variation - in μGal/yr and in Eötvös/yr – and in terms of their characteristic wavelength intervals, compared with the static GOCE and quasi static synthetic NGGM data.
In our study different active tectonic processes have been addressed, such as subduction, collision and extension, which can be correlated to earthquakes of different nature, thrust, normal and strike slip; a wide spectrum of tectonic settings, such as ocean-ocean or ocean continent subduction, subduction velocity, from slow to fast subduction, subduction dip, from shallow to deep subduction, as well as different age, from young to old lithosphere, have also been considered.
Our study aims to define the theoretical framework characterizing the Solid Earth processes which evolve over time up to the present-day density structure of the crust and upper mantle, so that the exploitation of NGGM data could help us to constrain the physics of the processes leading to the present-day structure of the Earth and its ongoing changes affecting our environment.


* The GravSeis Group comprises: R. Sabadini, A.M.Marotta, G. Cambiotti, A. Bollino (Department of Earth's Sciences, Università degli Studi di Milano, c/o L. Cicognara 7, 20129 Milan, Italy); C. Stefano, A. Anselmi (Domain Exploration & Science Italy, Thales Alenia Space Italia S.p.A, Torino, Italy); N. Sneeuw, K. Douch (Institute of Geodesy, University of Stuttgart, Stuttgart, Germany), Gravitational Seismology project, funded by the European Space Agency (ESA) as part of the Activity Line 1: Scientific Data Exploitation.