CICLO DE SEMINÁRIOS DO LABSIS – 12º SEMINÁRIO

Título: Bayesian Inversion of Receiver Functions and Surface Wave Dispersion Data from the Northeast of Brazil

Autora: Thuany Patrícia Lima

Data: 24 de novembro de 2016 – Horário: 15:30h
Local: Auditório do Departamento de Geofísica (2º pavimento do Prédio REUNI)

Resumo: We present crustal shear wave velocity models beneath 43
stations in the Northeast region of Brazil, using a method of Bayesian
joint inversion for receiver functions and surface wave dispersion
data. Our approach can be divided into two major steps combining the
estimation of the model parametrization by a trans-dimensional
optimization procedure and once it has defined the number of layers of
the model, a Bayesian sampling is performed for the uncertainty
estimation in a fixed dimension. Ideally, in a fully Trans-dimensional
Bayesian inversion, the level of data determines the complexity of the
model constraining the optimum number of unknowns (model parameters)
as well as the different levels of uncertainty brought by different
data types. However, it’s important to consider that this kind of
approach demands an extremely high cost of computational efficiency.
We avoid some of the computational burden involved in such procedure
by selecting an optimum model parametrization by the use of a
trans-dimensional algorithm of optimization, and then the sampling for
the uncertainty estimation is carried out with the settled
parametrization. Data errors are inferred by a combination of
empirical and hierarchical approaches. The inversion is performed to
observed data from the Borborema Province region (NE of Brazil), which
consists of a complex tectonic framework characterized by different
geological domains. Rayleigh wave group velocity dispersion curves and
receiver functions are jointly inverted to investigate the Moho depth
in the area, making use of a different methodology from what have been
used so far for that region. The results presented in this work are
coherent to previous published results that states the crust in the
Borborema Province region can be divided into two types: a thicker
crust (35-41km) observed in the high-standing topography of the south
of Borborema Plateau; and a thinner crust (29-34km) observed in the
low-standing topography of the province. The thickness of the crust is
overall well correlated to the topography, except for the north of the
Borborema Plateau which is a highly elevated region and the models
display consistently thin crustal thickness, similar to what was
reported for depression areas. Some authors interpret that the
depression areas and the north region of the plateau have experienced
a strong stretching during the Mesozoic, while the stronger
rheological south region has not. The present high elevation of north
is then related to an uplift that might have occurred during the
Cenozoic.