High-resolution lithospheric structure in the Gibraltar arc using P and S receiver functions
Resumen Abstract Índice Conclusiones
Molina Aguilera, Antonio
2021-A
Descargar PDF 
El límite de las placas euroasiática y africana en la zona más occidental del Mediterráneo se distribuye de forma irregular y configura un arco orogénico, el arco de Gibraltar. La región exhibe procesos tectónicos que incluyen subducción oceánica y continental, así como delaminación continental. Aunque la geodinámica de la zona está ampliamente interpretada de acuerdo a una subducción en retroceso hacia el este, la relación entre la dinámica de la laja que subduce y la tectónica más superficial es aún controvertida. En esta tesis aportamos luz sobre la compleja geodinámica de la región Íbero-Magrebí mediante la obtención de imágenes de alta resolución de la estructura cortical y del manto superior usando la técnica de las funciones receptoras de onda P y S.
Las imágenes sísmicas obtenidas con las funciones receptoras S a lo largo del arco de Gibraltar muestran la conexión mediante un corredor oceánico entre la laja de Alborán situada bajo el mar de Alborán y la corteza oceánica Atlántica. Interpretaciones de funciones receptoras previas revelaron la presencia de heterogeneidades laterales significativas en la litosfera sugiriendo procesos de cabalgamiento y delaminación continental en el sudeste de España. Analizamos dos perfiles sísmicos pasivos de alta resolución para dibujar la estructura cortical en la zona central y oriental del orógeno Bético, implementando una variedad de métodos de funciones receptoras que son complementarios: (1) la migración convencional de punto de conversión común (Common Conversion Point Stacking), basada en la hipótesis
simplificadora de modelo de capas plano-paralelas, que ha permitido iluminar la estructura cortical más importante a lo largo de los perfiles; (2) la migración de tiempo invertido (Reverse Time Migration), que emplea toda la complejidad de la forma de onda, generando imágenes más claras de la topografía de la discontinuidad entre la corteza y el manto (Moho); (3) la descomposición en armónicos de las funciones receptoras de onda P según el back-azimuth, que ha ayudado a constreñir la geometría e intensidad de las heterogeneidades laterales así como la obtención de una función receptora sólo sensible a la estructura plano-paralela del subsuelo, que puede (4) ser invertida conjuntamente con curvas de dispersión de ondas superficiales y así obtener un perfil promedio 1D de velocidad en profundidad.
Todos estos métodos arrojan resultados coherentes y complementarios. Uno de los resultados más importantes de este trabajo es la presencia de un salto cortical brusco y prominente (15-17 km) observado en ambos perfiles de alta densidad, que coincide con el límite entre los dominios de Alborán e Ibérico, y es interpretado como una falla de desgarro vertical (o STEP) cuya propagación hasta la superficie a lo largo de la placa que cabalga (dominio de Alborán) se produce en forma de estructura de flor positiva. Esta falla de desgarro acomodó las diferencias en la velocidad de la subducción en retroceso (roll-back) a lo largo de la dirección de subducción, en el flanco norte del sistema Mediterráneo occidental, cuando la litosfera continental Ibérica comenzó a subducir bajo el dominio de Alborán.
Los cambios en la topografía de la discontinuidad Moho, y la geometría de la falla de desgarro en la parte central de las Béticas han venido propiciados por debilidades en el paleomargen de Iberia. No se aprecian raíces corticales bajo las mayores altitudes en ambos perfiles, sugiriendo que la topografía más elevada se debe a la combinación de levantamiento producida por la estructura en flor positiva y el empuje de la astenosfera producido por el desgarro de la litosfera subducida de Iberia a lo largo de la falla de desgarro. Usando la descomposición en armónicos de las funciones receptoras de componente radial y transversal, se ha obtenido una representación continua de la heterogeneidad lateral en profundidad a lo largo de la corteza y el manto superior, determinando la orientación de la anisotropía y el buzamiento de las discontinuidades principalmente corticales. Este análisis distingue varias zonas con características propias a lo largo
de la región: (1) el dominio de Alborán localizado en el sudeste de Iberia, cerca de la línea de costa, se caracteriza por una Moho que buza en dirección NNO y una anisotropía regional gobernada por el esfuerzo regional (stress-induced shear) y anisotropía estructural en la zona cercana a las fallas principales de la zona de cizalla de las Béticas orientales (Eastern Betics shear zone), (2) la zona de transición entre el dominio de Alborán y el macizo Ibérico que coincide con la falla de desgarro está caracterizada por la presencia significativa de buzamiento y/o anisotropía a profundidad cortical y subcortical y (3) la región estable, sin presencia importante de heterogeneidad lateral asociada al macizo Ibérico.
Plate margin between Africa and Eurasia in westernmost Mediterranean is distributed as an irregular boundary that configures a tight orogenic belt, the Gibraltar arc. The region exhibits tectonic processes which may involve both oceanic and continental subduction and delamination. Thought the geodynamics of the zone has mainly been interpreted in terms of east dipping subduction rollback, the relationship between slab dynamics and surface tectonics is still under considerable debate. In this thesis we shed some light on the complex geodynamics of the Ibero-Maghrebian region by imaging precisely the crustal and upper mantle structure using the S and P receiver function approaches. The seismic imaging with S receiver functions all along the Gibraltar arc reveals the connection through an oceanic corridor between the Alboran slab under the Alboran Sea and the Atlantic oceanic crust. Previous P receiver function interpretations showed significant lateral heterogeneities in the lithosphere and suggest processes of underthrusting and continental
delamination occurring mainly in Southeastern Spain. We analyze two highresolution passive seismic profiles to probe the crustal structure in the central and eastern Betic orogen, implementing different and complementary receiver function interpretation techniques: (1) the conventional Common Conversion Point stacking method based on the simplifying flat-layered assumption that enhances the main crustal structure along the profiles; (2) the reverse time migration technique, which exploits the whole complexity of the wavefield, produces
clearer images of the Moho topography; (3) backazimuth harmonic decompositions of receiver functions to constrain the geometry and strength oflateral heterogeneities that also allows to obtain an only sensitive to the flatlayered structure receiver function to be (4) jointly inverted with surface wave dispersion curves and therefore obtain the average 1D S-wave velocity depth profile.
All these methods give coherent and complementary results. A sharp and prominent crustal step (15-17 km) is observed in both profiles, coinciding with the boundary between the Iberian and Alboran domains and interpreted as a near-vertical STEP fault that propagates to the surface as a positive flower fault structure. This STEP fault accommodated the differences in the subduction rollback velocity along the strike, at the northern edge of the Western Mediterranean system, when the thinned Iberian continental lithosphere started subducting under the Alboran domain. The changes in the topography of the Moho discontinuity, and the geometry of the STEP fault in Central
Betics is driven by inherited weaknesses in the Iberian paleomargin. No crustal roots are observed under the highest altitudes of both profiles suggesting that its high topography is due to a combination of the uplift produced by the positive flower structure and the push up of the asthenosphere after the lithospheric removal of the underthrusting Iberia along the STEP fault. Using the harmonic decomposition of radial and transverse receiver functions we obtain a continuous representation of the lateral heterogeneity with depth along the crust and the uppermost mantle, determining the orientation of anisotropy and dipping interfaces. This analysis distinguishes several zones along the region: (1) the Alboran domain located in the SE section nearby the coastline and mainly characterized by a dipping NNW Moho discontinuity and a regional anisotropy governs by stress-induced shear anisotropy and structural anisotropy nearby the main faults of the Eastern Betics shear zone, (2) the transition zone between the Alboran domain and the Iberian Massif coinciding with the STEP fault and marked by the presence of significant dipping and/or anisotropy at crustal and subcrustal depths and (3) a stable region without considerable lateral heterogeneity.
Contents
Abstract viii
Acknowledgements xi
1 Introduction 1
2 The Gibraltar arc tectonic setting 5
2.1 Plate tectonic setting . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Tectonic evolution of the western Mediterranean . . . . . . . . . 10
2.3 Geology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.1 External Zone or foreland thrust belt . . . . . . . . . . . . 14
2.3.2 Internal Zone or Alboran domain . . . . . . . . . . . . . 16
2.3.3 Alboran basin . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3.4 Gulf of Cadiz . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.3.5 Intramountain basins . . . . . . . . . . . . . . . . . . . . . 20
2.4 Seismicity and active tectonics . . . . . . . . . . . . . . . . . . . . 21
2.5 Current models and present-day lithospheric structure . . . . . 23
3 Data and Method 31
3.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.1.1 Gibraltar arc virtual seismic network . . . . . . . . . . . 34
3.1.2 High-resolution passive seismic profiles:
Hire and Transcorbe experiments . . . . . . . . . . . . . . 36
3.2 Receiver Function Method . . . . . . . . . . . . . . . . . . . . . . 39
3.2.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . 42
Conversion: Physics of the problem . . . . . . . . . . . . 42
Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Deconvolution . . . . . . . . . . . . . . . . . . . . . . . . 44
Single-station RF stacking and moveout correction . . . 47
3.2.2 Interpretation Methods . . . . . . . . . . . . . . . . . . . 49
Phase-weighted Common Conversion Point stacking migration
. . . . . . . . . . . . . . . . . . . . . . . 49
Reverse Time Migration . . . . . . . . . . . . . . . . . . . 53
Harmonic Decomposition Analysis . . . . . . . . . . . . 54
Joint Inversion . . . . . . . . . . . . . . . . . . . . . . . . 54
4 S-RFs migration in the Gibraltar arc system 55
4.1 Data and Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5 P-RF migration in high-dense profiles: Hire and Transcorbe 67
5.1 Data and Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.1 Hire profile . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.2 Transcorbe profile . . . . . . . . . . . . . . . . . . . . . . . 85
5.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6 Reverse Time Migration in the Hire seismograph 95
6.1 Data and Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.1.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.1.2 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
7 Harmonic Decomposition Analysis in Eastern Betics 119
7.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
7.1.1 Dependency of RFs on lateral heterogeneities . . . . . . . 122
7.1.2 Harmonic Decomposition Analysis . . . . . . . . . . . . 127
7.1.3 Illustration of the method . . . . . . . . . . . . . . . . . . 129
7.1.4 Testing the inclusion of both components in the Harmonic
Decomposition . . . . . . . . . . . . . . . . . . . . . . . . 134
7.2 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
7.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
8 Joint Inversion of RFs and surface wave dispersion curves in Eastern
Betics 159
8.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
8.1.1 Linearized Joint Inversion Method . . . . . . . . . . . . . 162
8.1.2 Harmonic Decomposition Analysis: Backazimuth independent
receiver function . . . . . . . . . . . . . . . . . . 164
8.2 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
8.2.1 Eastern Betics Zone . . . . . . . . . . . . . . . . . . . . . . 171
8.2.2 High-resolution Transcorbe profile . . . . . . . . . . . . . 181
8.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
9 Conclusion 191
A Harmonic Decomposition Figures 195
B Joint Inversion Figures 203
Bibliography 225
2. By migration high-resolution PRFs profiles we obtain the first direct images (not inverted) of the crustal and lithospheric architecture of the Betic orogen. From the interpretation of the CCP images we observe as the underthrusted Iberian lithosphere terminates sharply in both profiles following at the surface the External-Internal Zone limit. This sharp and prominent lithospheric step (15 km) observed in both profiles is interpreted as a near-vertical STEP fault that propagates to the surface as a positive flower fault structure. This STEP fault accommodated the differences in the subduction rollback velocity along the strike, at the northern edge of theWestern Mediterranean system, when the the thinned Iberian continental lithosphere started subducting under the Alboran domain.
We have not observed a crustal root under the highest altitude of both profiles, especially under Sierra Nevada mountain at the Hire profile,
suggesting that the high topography is produced by a combination of the uplift produced by the positive flower structure and the push up of
the asthenosphere after the lithospheric removal of the underthrusting Iberia along the STEP fault. Our results also present how the lithospheric tearing associated with a STEP propagation in a large scale subduction rollback framework can drive to high oblique continental collision which can be accommodated through crustal scale decollements and be a first order mechanism related with the built of fold and thrust belts of the peripheral External Betic Range.
3. We apply the Reverse Time Migration (RTM) methodology, which is able to exploit the complete complexity of the wavefield, to the teleseismic converted phases recorded in the high-density Hire seismic profile to create a high-resolution image of the crustal structure. RTM produces clearer images of the crustal structure and better resolution of lateral changes than CCP stacking. The main observations include a sequenceof abrupt Moho offsets, one of them associated with a STEP fault, and variations in Moho dip. We propose that the principal Moho topography features conform a deformed block of 60 km length formed by continental/ transitional Iberian crust which concentrates most of the deformation due to the collision. We suggest that the changes in the topography of the Moho discontinuity, and the geometry of the STEP fault in Central Betics is driven by inherited weaknesses in the Iberian paleomargin.
4. We implement a technique for the study of lateral heterogeneity employing RFs and apply it to a real teleseismic dataset in a region geologically diverse and structurally complex. The technique permits to separate the isotropic information from the energy due to anisotropic and/or dipping subsurface structures, becoming a useful tool for the understanding of subsurface geometries. The method successfully use both radialand transverse components which gives better constraints on the subsurface seismic structure than radial-only RFs analysis. The inclusion of the transverse component helps to determine if the lateral heterogeneity is due to dipping interfaces and/or anisotropic layers as the transverse component is mainly sensitive to these sources of lateral heterogeneity.
The harmonic decomposition method implemented is linear and allows to build a time dependent vector representation of the terms in the expansion, simplifying notably the interpretation. The alignment of the vectors ~a1 and ~a2 and the time delays at which they are concentrated help to remove some ambiguities inherent to the kind of lateral heterogeneity we Through a synthetic study we illustrate the method. We investigate three basic scenarios; a dipping interface, a flat anisotropic layer and an inclined anisotropic layer. Furthermore, we evaluate the goodness of the moveout correction on radial and transverse components and conclude that the transverse component is less sensitive to the slowness variation of RFs. This, combined to the fact that transverse RFs are mainly affected by the lateral heterogeneity due to anisotropic layers and dipping interfaces, permit through comparative radial- and transverse-only harmonic decomposition approaches determine with less ambiguity the source of the lateral heterogeneity.
We apply the harmonic decomposition method using a comparative radial and transverse component study to real RFs located in the southeastern region of the Iberian Peninsula. We obtain a continuous representation of the lateral heterogeneity with depth along the crust and the uppermost mantle, determining the orientation of anisotropy and dipping interfaces.
Our study shows that crustal and uppermost mantle anisotropic characteristics might differ in their orientation. Stations located nearby
the coastline and faults exhibit the presence of considerable lateral heterogeneity at crustal depths. The vector ~a1 is oriented predominately in a NNW direction, in agreement with previous results which report a gently dipping Moho following in this direction (Fig. 7.11B). The vector ~a2 seems to be geometrically related to the fault system present in the region and related to the vector ~a1 at crustal depths. At deeper depths, below the Moho, the vector ~a2 reaches the largest lengths for the stations located above a region with double Moho. This anisotropy might be related with an anisotropic structure trapped between the two Moho discontinuities.
5. We demonstrate the effectiveness of the harmonic decomposition analysis to isolate the signal in RFs only sensitive to the flat-layered isotropic structure of the medium. This analysis permits to obtain an independent of the backazimuth RF that can be used to invert the 1D S-wave velocity-depth profile. This approach is a clear advantage in presence of significant lateral heterogeneity associated to dipping and/or anisotropic layers compared to conventional schemes where RFs grouped in backazimuth ranges are inverted (BAZ approach). We theoretically prove this performing synthetic inversion experiments with different lateral heterogeneity scenarios.
Real joint inversions of RFs and surface wave dispersion measurements comparing both approaches are as well made in the Eastern Betics zone
where different types of lateral heterogeneity are present. The harmonic decomposition approach is able to capture accurately intracrustal and
crust-mantle discontinuities. Compared to the BAZ approach, the HD one especially succeeds in regions where the lower crustal structure is
considerably deformed, as happens along the STEP fault where a double Moho jumps is recovered.
The goodness of the HD scheme is also proved with the high-resolution Transcorbe transect. The 1D shear-wave velocity cross-section along the
profile obtained with the HD approach retrieves the main crustal and upper-mantle features revealed with the CCP stacking migration; Moho
offset and low velocity zone. Furthermore, the combined use of the inversion and theHDmethod permits the projection in depth of the lateral heterogeneity geometry. This analysis distinguishes several zones along the profile in coherence with the previous results: (1) the Alboran domain located in the SE section nearby the coastline and mainly characterized by a dipping NNW Moho discontinuity, (2) the transition zone between the Alboran domain and the Iberian Massif coinciding with the STEP fault and marked by the presence of significant dipping and/or anisotropy at crustal and subcrustal depths and (3) the stable Iberian massif region without considerable lateral heterogeneity.