Tomographic models of seismic velocity perturbations provide valuable insights into the Earth's interior. Yet, because seismic velocities are primarily temperature-dependent, they mostly capture thermal anomalies. Combining them with gravity data enables a joint interpretation that can also account for compositional variations in the upper mantle beneath Precambrian cratons. The South African cratonic region is composed of the Archean Kaapvaal and Zimbabwe craton and, according to xenolith analysis, underlain by a heterogeneous upper mantle both in terms of temperature and composition. To unravel the temperature and compositional anomalies with corresponding density variations and to link them to the tectonic history of that area, we apply an integrative technique based on a joint interpretation of the seismic tomography and gravity data. We combine a global shear seismic tomography model with an embedded high-resolution regional model and we invert it for temperature, assuming an initial composition, representative of a refertilized upper mantle. The composition and temperature of the upper mantle are iteratively changed, increasing progressively the amount of iron depletion, to fit the residual density, obtained from the joint inversion of the residual gravity and residual topography, and GOCE gravity gradients. The results show significant lithospheric compositional variations consistent with the tectonic history of the area. The most depleted lithosphere is located in the Southeastern Terrane of the Kaapvaal craton, at depth < 100 km, generating a temperature higher than similar to 150 degrees C, with respect to that of a refertilized lithosphere.

Xenolith data reveal lateral and vertical compositional variations of the upper mantle of the Precambrian cratons, indicating a different degree of refertilization with respect to the most depleted mantle in iron components, characterizing the oldest Archean cratons. The South African cratonic region is composed of the Kaapvaal and Zimbabwe craton, both of Archean age, having deep and fast lithospheric roots, which are likely depleted in heavy constituents. In contrast, there exist regions, such as the Limpopo belt, a terrane that was trapped between the Kaapvaal and Zimbabwe cratons during their collision (2.6–2.7 Ga), and Bushveld Complex, an area characterized by intraplate magmatism occurred 2.05 Ga, whose negative velocity anomalies in the upper mantle, indicate a more fertile composition due to metasomatism. To unravel the origin of these anomalies and link them to the tectonic history of the area, we apply an integrative technique based on a joint interpretation of the seismic tomography and gravity data, which can discern temperature and compositional variations. To this aim, we combine the global surface seismic tomography model [1] with the embedded regional model [2], derived from teleseismic tomographic inversion of the S-body wave dataset recorded by the Southern African Seismic Experiment. The combined seismic model is inverted for temperature, assuming an initial composition, representative of a refertilized upper mantle [3], using a mineral physics approach [4]. The composition and temperature of the upper mantle are iteratively changed, increasing progressively the amount of iron depletion, to fit the residual density, obtained from the joint inversion of the residual gravity and residual topography. The great advantage of using both the gravity field and residual topography lies in their different dependence on the distribution of density heterogeneities (depth and size). In a second type of inversion we included the GOCE gravity gradient [5]. The obtained results show that the most depleted lithosphere is confined at depth lower than 100 km, generating a temperature higher than ~200, with respect to that of a refertilized lithosphere. The Southeastern Terrane of the Kaapval craton are characterized by thicker and more depleted cratonic roots than the Zimbawe craton. The presence of a depleted mantle below the cratonic crust may indicate that the crust and mantle have been connected since the craton formation. These results, related to the different structures and properties of the upper mantle, improve our understanding of the evolution of the South African cratonic lithosphere. References [1] Schaeffer and Lebedev, 2013. https://doi.org/10.1093/gji/ggt095 [2] Youssof et al., 2015. http://dx.doi.org/10.1016/j.epsl.2015.01.034 [3] Griffin et al., 2004. doi:10.1016/j.chemgeo.2004.04.007 [4] Conolly, 2005. doi:10.1016/j.epsl.2005.04.033 [5] Kaban et al., 2022. doi.org/10.1007/s00024-021-02925-

Special thanks to Derek Keir, Finnigan Illsley-Kemp (University of Southampton), Jose A. Lopez Comino (KAUST and now at University of Granada), and Kiran K. Thingbaijam (KAUST) for the fruitful discussions which include suggesting FI technique and enhancing the moment tensor inversion and stress drop analyses. We wish to thank Felix Waldhauser (Lamont Doherty Earth Observatory) for providing the latest unpublished version of his code (HypoDD of 2019). Our project received financial support from King Abdullah University of Science and technology (KAUST) under the CRG grant (ORS-2016-CRG5-3027-0). All waveform data were obtained through the Saudi Geological Survey (SGS) and is available from Harvard Dataverses webpage (https://doi.org/10.7910/dvn/kw4zfx).

Magmatic intrusions often produce ground deformation that can be studied by geodetic techniques. In the past two decades, many volcanic dike and sill emplacements (sometimes associated to eruptions) in different tectonic settings have been analyzed through InSAR. However, in only a few cases, the post-intrusive behavior has been studied. Here we analyze the post-diking deformation in Harrat Lunayyir, which is a mononogenetic volcanic field located in western Saudi Arabia on the eastern margin of the Red Sea Rift.Between April and July 2009, an intensive seismic swarm occurred in the area with many earthquakes above magnitude 4 and the largest earthquake of Mw 5.7. InSAR data showed that the earthquake swarm was triggered by the emplacement of a dike intrusion that stopped only ~1 km below the surface. Dike length was estimated to be ~7 km and with a maximum opening 4 m. Above the intrusion, a ~10 km long and ~5 km wide graben formed during the activity with up to 1 m of fault slip on the border faults. In the post diking phase up to present, micro-sesmicity (0In February 2017, a new seismic swarm occurred ~60 km north of Harrat Lunayyir and another swarm started in October 2018, about 30 km southwest of the volcanic field. Both swarms are still ongoing with a few events per week and Ml<3.5. By using Sentinel-1 images, acquired during the period 2015-2019, we derived deformation rate maps for the entire Harrat Lunayyir volcanic field. No ground deformation was detected at the locations of the recent seismic swarms, and a thorough analysis of seismic signals excludes the swarms were caused by new magmatic intrusions. However, within the Harrat Lunayyir graben region, we noticed a steady and long-lasting subsidence of ~1 mm/yr. During the 2015-2019 period, the total seismic moment release would only be able to accommodate less than 0.1 mm of the observed subsidence and thus the current post-diking deformation is mainly aseismic.In order to reconstruct the entire post-diking deformation history in Harrat Lunayyir we also analyze older available SAR images (Envisat, ALOS, TerraSAR-X, TanDEM-X). Our preliminary results show that the subsidence rate in the graben area was faster just after the intrusion (few cm in two months) but then rapidly decayed as well as the seismicity. We are now investigating different processes that can cause post-diking deformation, such as residual opening of the dike, post-diking settlement of faults and fractures, release of gases into fractures, cooling of the dike, and post-diking viscoelastic relaxation. Modelling of the deformation source will contribute to the understanding on which of these post-diking processes might be the dominant one in Harrat Lunayyir.

Are recent seismic swarms Lunayyir Arabia) associated with magmatic intrusions? Copernicus/EGU. In volcanic fields, magmatic intrusions are usually preceded and accompanied by seismic swarm. Even if the intrusions do not culminate in an eruption, these events typically produce ground deformation that can be observed using geodetic techniques such as InSAR. It is therefore important to combine seismic and geodetic data to better understand these magmatic processes and to find out whether a seismic swarm is associated with a magmatic in-trusion or not. One example occurred in Harrat Lunayyir volcanic field, in Saudi Arabia, where a seismic swarm of ∼ 30000 earthquakes hit between April and July 2009 with the largest earthquake of Mw 5.7. Extensive surface fractures formed over 10km in length at the height of the activity. The ground displacements measured with InSAR showed that the swarm was accompanied by a dyke intrusion, estimated to be ∼ 7km in length and with an opening of up to 4m. During the past two years, the Harrat Lunayyir region has been hit by two new seismic swarms. The first one, located about 60km to the North, started in February 2017 and lasted until August 2018. The second swarm, located about 30km to the West, started in late September 2018 and is still ongoing. These swarms are characterized by frequent earthquakes of Ml<4. To assess if significant ground deformation was associated with the swarms, we used InSAR technique again. We processed Sentinel-1 SAR images acquired from both ascending and descending orbits between January 2017 and December 2018. Due to the general high co-herence, the interferograms could easily be unwrapped and used to calculate deformation rate maps in the line of sight (LOS) of the satellites with the SBAS technique. The resulting rate maps showed signals correlated to topography, indicating that significant elevation-related atmospheric delays were affecting the results. We reduced this signal by estimating linear correlation coefficients between elevation and the LOS signal and subtracted the results from the rate maps. The corrected deformation rate maps mostly show spatially smooth LOS-rate variations of [U+F0B1] 0.5cm/year, which are due to strong atmospheric delays in some of the interferograms that could not be fully removed in the processing. No detectable deformation is seen at the locations of the relocated seismic swarms. Using the relocated hypocenters of the first seismic swarm, we estimated the fault geometry and the likely total fault slip and roughly quantify the expected ground deformation. The maximum surface displacement associated to slip on a 4km x 8km normal fault dipping 30 ◦ at 5km depth is less than 0.5mm and thus not detectable with InSAR. In contrast, a sub-vertical 4km x 8km dyke with 0.5m opening at the same depth would produce more than 1cm of ground deformation. Since no clear ground deformation was detected in the two areas affected by the seismic swarms, we conclude that they were not accompanied by shallow magmatic intrusions. However, the SGS will continue to monitor the area with their permanent seismic network and the InSAR deformation maps will be updated to detect any anomalous ground deformation.

The long-term stability of Precambrian continental lithosphere depends on the rheology of the lithospheric mantle as well as the coupling between crust and mantle lithosphere, which may be inferred by seismic anisotropy. Anisotropy has never been detected in cratonic crust. Anisotropy in southern Africa, detected by the seismological SKS-splitting method, usually is attributed to the mantle due to asthenospheric flow or frozen-in features of the lithosphere. However, SKS-splitting cannot distinguish between anisotropy in the crust and the mantle. We observe strong seismic anisotropy in the crust of southern African cratons by Receiver Function analysis. Fast axes are uniform within tectonic units and parallel to SKS axes, orogenic strike in the Limpopo and Cape fold belts, and the strike of major dyke swarms. Parallel fast axes in the crust and mantle indicate coupled crust-mantle evolution for more than 2 billion years with implications for strong rheology of the lithosphere.

This study presents seismic images of the crustal and lithospheric structure in Siberia based on the available broadband seismic data using teleseismic receiver functions (RFs). We invert P- and S-RFs jointly. The inversion technique is carried out by approach described by Vinnik et al. (2004). With this method, we determine seismic P- and S-velocities that are comparable to the results of teleseismic body wave and surface wave tomography techniques.TheRFmodel shows variations in the crustal thickness between 35 and 55 km. Intracrustal structures are identified, in particular using the high-frequency P-RF component as it has about an order of magnitude better resolution than S-RF. We find no indication for significant crustal anisotropy in the cratonic areas of Siberia. The preliminary crustal thickness results from the Hk stacking and from the inversion approach agree with a previous study of mainly controlled source results by Cherepanova et al. (2013). Here we also determine the Vp/Vs ratio. Our analysis maps deep features in the upper mantle as the mid-lithospheric discontinuity and the thickness of the transition zone, whereas the lithosphere-asthenosphere boundary is not resolved. The current results of RF analysis of the crustal and mantle structure will help to build a model for tectonic and geodynamic evolution of different provinces of Siberia. We compare our results to the recent detailed of crustal structure in area and with models for similar geodynamic

Earlier seismic studies of the Kalahari Craton in southern Africa infer deformation of upper mantle by flow with fast direction of seismic anisotropy being parallel to present plate motion, and/or report anisotropy frozen into the lithospheric mantle. We present evidence for very strong seismic anisotropy in the crust of the Kalahari craton, which is 30-40% of the total anisotropy as measured by SKS splitting. Our analysis is based on calculation of receiver functions for the data from the SASE experiment which shows strong splitting between the SV and SH components. The direction of the fast axes is uniform within tectonic units and parallel to orogenic strike in the Limpopo and Cape fold belts. It is further parallel to the strike of major dyke swarms which indicates that a large part of the observed anisotropy is controlled by lithosphere fabrics and macroscopic effects. The directions of the fast axes for the crustal anisotropy are parallel to the general directions determined from SKS splitting, although the directions from our analysis of receiver functions is more homogeneous than for SKS splitting. This analysis indicates parallel fast axes in the crust and in the mantle, which suggests that the crust and lithospheric mantle may have been coupled since cratonisation. If so, the apparent match between mantle anisotropy and the present plate motion is coincidental.

We present a 3D high-resolution seismic model of the southwestern Africa region from teleseismic tomographic inversion of the P- and S- wave data recorded by the amphibious WALPASS network. We used 40 temporary stations in southwestern Africa with records for a period of 2 years (the OBS operated for 1 year), between November 2010 and November 2012. The array covers a surface area of approximately 600 by 1200 km and is located at the intersection of the Walvis Ridge, the continental margin of northern Namibia, and extends into the Congo craton.

We present a 3D high-resolution seismic model of the southern African cratonic region from teleseismic tomographic inversion of the P- and S-body wave dataset recorded by the Southern African Seismic Experiment (SASE). Utilizing 3D sensitivity kernels, we invert traveltime residuals of teleseismic body waves to calculate velocity anomalies in the upper mantle down to a 700 km depth with respect to the ak135 reference model. Various resolution tests allow evaluation of the extent of smearing effects and help defining the optimum inversion parameters (i.e., damping and smoothness) for regularizing the inversion calculations.The fast lithospheric keels of the Kaapvaal and Zimbabwe cratons reach depths of 300-350 km and 200-250 km, respectively. The paleo-orogenic Limpopo Belt is represented by negative velocity perturbations down to a depth of 250 km, implying the presence of chemically fertile material with anomalously low wave speeds. The Bushveld Complex has low velocity down to similar to 150 km, which is attributed to chemical modification of the cratonic mantle. In the present model, the finite-frequency sensitivity kernels allow to resolve relatively small-scale anomalies, such as the Colesberg Magnetic Lineament in the suture zone between the eastern and western blocks of the Kaapvaal Craton, and a small northern block of the Kaapvaal Craton, located between the Limpopo Belt and the Bushveld Complex. (C) 2015 Elsevier B.V. All rights reserved.