[1] We reconstruct the subsurface geology in a region of the northern Apennines (central Italy) where a protracted extensional sequence occurred in 1997-1998 with maximum magnitude M = 6.0. Our study is mainly based on the interpretation of three reprocessed seismic reflection profiles crossing the epicentral area, which constrain the subsurface geometry to a depth of about 12 km where most of the shallow seismicity occurs. Comparing the subsurface setting with accurately determined earthquake locations, we find that the seismicity is located entirely within the sedimentary cover and does not penetrate the underlying basement. This is explained by considering that the sedimentary cover is rather thick and composed of relatively strong lithologies (platform carbonates and evaporites), while the upper part of the basement consists of weak phyllites and siliciclastic rocks. This weak horizon is also evidenced by the low-Vp values measured in deep wells of the region. Its effect is to decouple the sedimentary cover from the crystalline basement, where only microseismicity occurs. Our study indicates that local structure and stratigraphy can significantly influence the distribution of seismicity within the upper crust, particularly in complex geological environments such as thrust-and-fold belts. Citation: Mirabella, F., M. Barchi, A. Lupattelli, E. Stucchi, and M. G. Ciaccio (2008), Insights on the seismogenic layer thickness from the upper crust

We discuss a data‐processing sequence adopted to reprocess a seismic line that crosses the Italian southern Apennines from the Tyrrhenian Sea to the Adriatic margin and investigate both the overthrust and foreland areas. We first determine the main causes of the very low S/N ratio in the field data and then propose a processing sequence aimed at exploiting the signal content, also making use of a priori geological knowledge of this area. Our work indicates a combination of causes for the very low quality of the seismic data. These include length of the spread (about 20 km) that is unfavorable because of the rapid variation in the near‐surface geology, tectonic complexity, crooked‐line acquisition, and the rough topography associated with outcropping rocks characterized by highly variable velocities. Based on the outcome of this data analysis, we present a processing sequence driven by knowledge of the regional tectonic setting and by knowledge of the shallow subsurface geology. The main effort is in removing the large, near‐surface related noise components. The low S/N ratio makes it impossible or nearly impossible to successfully apply highly sophisticated techniques such as depth migration or wave equation datuming. Thus, we used robust techniques specifically designed to solve each problem that degraded data quality. The most relevant of these techniques were the removal of bad traces where unacceptably low quality was detected by energy and frequency decay criteria; estimation and correction for static time shifts attributable to near‐surface conditions; optimization of common midpoint (CMP) sorting to attenuate the deleterious effects of the crooked‐line acquisition; application of a weighted stacking technique to maximize stack power and application of prestack f-x deconvolution to attenuate uncorrelated noise. The outcome of this processing sequence is compared with the result of a more standard sequence that was previously applied to the same data and is also discussed in terms of the possible geological model it might evidence. The realization of a seismic section showing rather continuous and structured events down to 8 s which, depending on the interpretation, may be related to Moho discontinuity or to very deep sedimentary layers supports the efficacy of the processing approach we propose.

No abstract

We present a stochastic full-waveform inversion that uses genetic algorithms (GA FWI) to estimate acoustic macro-models of the P-wave velocity field. Stochastic methods such as GA severely suffer the curse of dimensionality, meaning that they require unaffordable computer resources for inverse problems with many unknowns and expensive forward modeling. To mitigate this issue, we propose a two-grid technique, that is, a coarse grid to represent the subsurface for the GA inversion and a finer grid for the forward modeling. We applied this procedure to invert synthetic acoustic data of the Marmousi model. We show three different tests. The first two tests use as prior information a velocity model derived from standard stacking velocity analysis and differ only for the parameterization of the coarse grid. Their comparison shows that a smart parameterization of the coarse grid may significantly improve the final result. The third test uses a linearly increasing 1D velocity model as prior information, a layer-stripping procedure, and a large number of model evaluations. All the three tests return velocity models that fairly reproduce the long-wavelength structures of the Marmousi. First-break cycle skipping related to the seismograms of the final GA-FWI models is significantly reduced compared to the one computed on the models used as prior information. Descent-based FWIs starting from final GA-FWI models yield velocity models with low and comparable model misfits and with an improved reconstruction of the structural details. The quality of the models obtained by GA FWI + descent-based FWI is benchmarked against the models obtained by descent-based FWI started from a smoothed version of the Marmousi and started directly from the prior information models. The results are promising and demonstrate the ability of the two-grid GA FWI to yield velocity models suitable as input to descent-based FWI

No abstract