A methodological comparison between low-frequency and high-frequency seismic reflection data for studying near-surface faults

Fault analysis in marine seismic data is conducted across various contexts, including tectonic, seismological and basin analysis studies, as well as in oil exploration and engineering projects. Generally, the higher the dominant frequency (F_dominant) of seismic data, the greater its vertical and horizontal resolution, making seismic features more representative of the geological record. However, most geological fault studies in marine environments rely on low-frequency seismic data (F_dominant ~ 50 Hz). As a result, geological records of deformation, erosion, and deposition at scales smaller than 8 meters remain invisible to interpreters, potentially leading to inaccurate structural interpretations. Despite this, the literature lacks comparative studies using real (i.e. non-modeled) data to assess the impact of seismic frequency on the concealment and/or distortion of geological features. This raises the following question: What kind of information in the geological fault record could be omitted from seismic interpretation when the seismic frequency is reduced?  This study employs both conventional (airgun source; F_dominant ~ 50 Hz, vertical resolution ~ 8 m) and high resolution (sparker source; F_dominant ~ 500 Hz, vertical resolution ~ 50 cm) multichannel seismic sections, which overlap the same fault that deforms the seafloor, to explore differences in the interpretation of its growth history. The normal fault analyzed has a minimum Quaternary age and is located above a salt dome in the Santos Basin (Southeast Brazil). A total of 36 seismic units were mapped in the sparker section, while the airgun visibility limit allowed only 13 units to be identified within the same stratigraphic interval (first 200 meters below the seafloor). Analysis of Throw-Depth Plots (T-D Plots) and Expansion Index (EI) revealed that the fault experienced 6 growth periods and 6 blind periods in the sparker data, while only 3 growth periods and 3 blind periods were identified in the airgun section. Only 65% of the growth and blind periods were synchronous between the two datasets. The sparker section revealed that noise features in the airgun data correspond to normal drags that generate footwall anticlines, hanging wall synclines, and synthetic dips in the fault's hanging wall. All seismic reflectors in the airgun section were plane-parallel. In contrast, the presence of offlaps, toplaps, and downlaps in the sparker data suggests that 4th and 5th order Quaternary sedimentary processes interacted with deformational features, generating differential thicknesses between footwall and hanging wall strata after fault growth periods. In summary, the comparative analysis demonstrated that reducing seismic frequency can result in: 1) underestimating the number of fault reactivation and quiescence periods; 2) hide ductile structures of shallow faults in marine sediments; and 3) suppress the identification of sedimentary processes that interacted with deformation features. Furthermore, the analysis of the high resolution seismic data shows that: 1) well-established fault analysis methods such as T-D Plots and EI should be analyzed together with stratigraphic features to avoid misinterpretations of growth periods; 2) it provides unprecedented level of detail about the Quaternary polycyclic evolution of a fault related to the halokinesis in the Santos Basin.