Seismic Imaging of the Subducted Australian Continental Margin Beneath Timor and the Banda Arc Collision Zone

Abstract

Detailed crustal and uppermost mantle structure is imaged for the first time utilizing ∼4 years of broadband seismic data newly collected in the Timor‐Leste and Nusa Tenggara Timor region of Indonesia. We apply three techniques, ambient noise tomography, teleseismic P wave receiver function, and coda autocorrelation, to resolve a 3D Vs model and Moho structure. Our tomographic images show low‐velocity anomalies (<30 km) beneath Timor related to underthrusted Gondwana sequence from the Australian plate, which are vertically offset by the high‐velocity backstop of the Banda forearc terrane. The structure progressively changes along strike, reflecting different collisional stages developed as a result of the oblique convergence. At greater depth, we detect seismically fast lithospheric mantle (>30 km) and the arc‐ward dipping Moho beneath Timor, both interpreted to be from the Australian plate. Our findings provide direct structural evidence of the Australian continental margin at lithospheric depths beneath the Banda Arc collisional zone.

A Detailed Earthquake Catalog for Banda Arc–Australian Plate Collision Zone Using Machine-Learning Phase Picker and an Automated Workflow

“…It is also approximately concordant with a geodetic block boundary (e.g., Nugroho et al, 2009), with the Timor-Wetar block in the east accommodating up to 63% motion of the Australian plate relative to southeast Asia compared to an amount of 21%-41% in the west (Nugroho et al, 2009). A recent ambient noise tomography study also observes contrasting crustal structure on the two sides, with more slower materials observed in the crust beneath the Timor (Porritt et al, 2016;Zhang and Miller, 2021;Miller et al, 2021). Combined, we interpret the abundant crustal seismicity beneath the Alor-Timor segment to reflect the strong and complex crustal shortening and subduction of Australian continental materials.…”

“…In total, EQTransformer yielded about 1.26 million P and 1.57 million S picks, which are then passed into step two, which uses the Rapid Earthquake Association and Location algorithm (REAL; Zhang et al, 2019) for association and preliminary earthquake location. We use a local 1D model derived from recent regional 3D P-and S-velocity models (Supendi et al, 2020;Zhang and Miller, 2021) (see supplemental material for details). We find that the outputs from REAL are mostly sensitive to the threshold parameter set for the number of associated phases.…”

Tectonic Fabric in the Banda Arc‐Australian Continent Collisional Zone Imaged by Teleseismic Receiver Functions

“…More interestingly, we recognize clear regional patterns in terms of the strikes from the A1 max signals as we show in Figure 4b. Together with existing knowledge of the crustal velocity structures by Zhang and Miller (2021) and inferences from the literature detailing geology and volcanism (e.g., Harris, 2011;Harris et al, 2009;Kaneko et al, 2007;Muraoka et al, 2002a), we cluster these RF results into four different areas (gray ellipses in Figure 4b). They are referred to as Zone A, B, C, and D hereafter, which are used to guide the discussion in this section.…”

“…The depth is determined from the delay time of those considered arrivals (vertical green lines in Figure 2) and station‐specific 1‐D models. Each 1‐D model was derived from a three‐dimensional (3‐D) crustal Vs model based on ambient noise tomography using the YS network data from Zhang and Miller (2021).…”

“…The pronounced along‐strike structural variations at shallow lithospheric depths may either result from the diachronous (progressive) collision as a result of the oblique convergence (Porritt et al., 2018; Zhang & Miller, 2021) or from inherent structural heterogeneities of the incoming and colliding Australian (lower) plate (Miller et al., 2021), or both. The present‐day seismic structure may be further complicated by exotic terranes or microplates formed during the Jurassic breakup of eastern Gondwana (e.g., Supendi et al., 2020; Zhang & Miller, 2021). Most recently, Jiang et al.…”