dc.description.abstract
Madagascar and Sri Lanka are key regions to understand the amalgamation and break-up of Gondwana; both islands were located centrally in the supercontinent and in the Pan-African Orogen. We determined crustal radial seismic anisotropy in southern Madagascar and the crustal velocity structure in Sri Lanka and compared our findings to those derived in other studies of Pan-African regions. We found major similarities, supporting the hypothesis that these once juxtaposed regions experienced major crustal unification through orogenetic processes.
We determined radial anisotropy (RA) in the crust of southern Madagascar from the differences between the speeds of vertically and horizontally polarized shear waves (VSV and VSH). The latter we derived from Rayleigh and Love surface wave dispersion determined from seismic ambient noise cross-correlations. The amalgamated Precambrian units in the east and the Phanerozoic Morondava basin in the west of southern Madagascar were shaped by different geodynamic processes: The crystalline basement was strongly deformed and metamorphosed to varying degrees during the assembly of Gondwana and the Pan-African Orogeny, whereas the Morondava basin was formed during the separation of Africa and Madagascar. The different developments are reflected in first order differences in the radial anisotropy structure.
In the Precambrian domains, positive RA (VSV<VSH) is found in the upper and lower crust, with a layer of negative RA (VSV>VSH) in between. The upper crustal anisotropy may reflect shallowly dipping layering within the Archean units and adjacent imbricated nappe stacks, whereas the lower crustal anisotropy likely represents fossilized crustal flow during the syn- or post-orogenic collapse of the Pan-African Orogen. The layer of negative RA may have preserved vertically oriented large shear zones of late Pan-African age. Within the Morondava basin, negative RA in the uppermost ∼5 km could have been generated by steep normal faults, jointing, and magmatic dike intrusions. The deeper sediments and underlying crustal basement are characterized by positive RA. This is consistent with horizontal bedding in the sediments and with the alignment of fabrics in the basement created by extension during the basin formation.
The crust of Sri Lanka mostly consists of Precambrian rocks. We analyzed newly derived data from a temporary seismic network deployed in 2016–2017. Rayleigh wave phase dispersion from ambient noise cross-correlation and receiver functions were jointly inverted for the seismic structure using a transdimensional Bayesian approach.
We determined Moho interface depths between 30–40 km, with the thickest crust (38–40 km) beneath the central Highland Complex (HC). The thinnest crust (30–35 km) is along the west coast, which experienced crustal thinning through the formation of the Mannar basin. The majority of VP/VS ratios are within a range of 1.66–1.73 and redominantly favor a felsic bulk crustal composition with intermediate to high silica content. A major intra-crustal (18–27 km), westward dipping (∼4.3°) interface with high VS (∼4 km/s) underneath is prominent in the central HC, continuing into the eastern Vijayan Complex (VC). The dipping discontinuity anda low velocity zone in the upper crust of the central HC can be related to the HC/VC contact zone. They are also in agreement with the well-established amalgamation theory of Sri Lanka, described by a stepwise collision of the arc fragments during Gondwana’s assembly, and deep crustal thrusting processes and a transpressional regime along the suture between HC and VC.
We found striking similarities of seismic properties between southern Madagascar and Sri Lanka, and southern India and East Antarctica as once juxtaposed Pan-African terranes. Their crustal thicknesses range between 35–40 km in regions with little to no influence of post-orogenic processes such as rifting or younger orogeny. Pan-African crust of all comparison studies agree to average crustal VP/VS ratios on the lower side (1.65–1.78), indicating felsic to intermediate bulk crustal compositions. Sri Lanka shows higher surface VS (3.1–3.6 km/s) compared to average values in southern Madagascar (3.2–3.4 km/s) and southern India (3.2 km/s), which might be due to different burial depths and metamorphic grade of surface rocks during the Pan-African Orogeny, or different petrological compositions. A low velocity layer in the upper crust as we observe in central Sri Lanka is also present in other Pan-African terranes, and possibly was generated through retrograde metamorphic processes and fluid migration during the transpressive regime. The strong similarities of seismic crustal properties between these regions might be attributed to the region spanning unification of crustal structures through extensive common overprinting during the Pan-African Orogeny. Differences might be seen as consequence of autochtone terrane compositions, positions within the orogen, and individual reworking processes after the orogeny.
We developed a software (BayHunter), implementing a Bayesian inversion approach, to estimate a range of models that fit the data and outputs meaningful uncertainty estimations. BayHunter is a Python framework to perform a Markov chain Monte Carlo (McMC) transdimensional Bayesian inversion of surface wave dispersion and receiver functions. The algorithm follows a data-driven strategy and solves for the velocity-depth structure, the number of layers, noise scaling parameters and VP/VS ratio. BayHunter was developed and used for the inversion of seismic data in Sri Lanka
en
