Analysis of Predictive and Spiking Deconvolution Effects on Seismic Image Quality of PSTM and CRS Sections in the Wetar Waters

Seismic image quality in structurally complex marine settings is strongly influenced by the choice of migration workflow and deconvolution operator. This study evaluates the performance of predictive and spiking deconvolution applied to 2D marine seismic data from the Wetar Waters using pre-stack time migration (PSTM) and common reflection surface (CRS) stack migration. The assessment is based on visual analysis of reflector continuity, amplitude behavior, and structural coherence along two seismic profiles: Line 1.3 and Line 14.1. Predictive deconvolution generally produces smoother, more laterally continuous reflectors, particularly when combined with CRS processing, whereas spiking deconvolution enhances wavelet sharpness but often results in less stable amplitude patterns in zones affected by strong structural deformation. The relative performance of each workflow varies between profiles. On Line 1.3, CRS with predictive deconvolution does not consistently outperform predictive PSTM in reflector definition, while on Line 14.1 the CRS-predictive combination provides clearer reflector continuity and more laterally coherent structural imaging, especially across folded and faulted intervals. These results indicate that no single processing sequence is universally superior. Instead, seismic image quality depends on the interaction between migration method, deconvolution strategy, and local geological complexity. Although the CRS workflow combined with predictive deconvolution in the Wetar waters generally improves reflector continuity and reduces noise sensitivity, the resulting sections exhibit smoother and relatively lower-frequency seismic character, which may reduce the visibility of subtle geological structures. Consequently, CRS-based imaging is more suitable for regional structural and stratigraphic interpretation, whereas finer structural details may be better preserved in PSTM sections.