Abstract: In exploration of ultra-deep small-scale karst fractured-vuggy reservoirs,conventional three-dimensional seismic data often have overly large bin sizes,which makes it difficult to identify effective signals such as diffraction and scattering waves,and cannot meet the requirements for refined characterization of fractured-vuggy bodies.Therefore,seismic data fusion processing technology was applied.Based on two three-dimensional seismic acquisitions in the Aiding 6 Wellblock of the Tahe Oilfield,geophysical attribute differences between datasets acquired in different years were systematically analyzed,including polarity,phase,time shift,frequency,energy,azimuth,and bin grids.Bin-normalization and offset-fusion processing were applied to develop a selective fusion processing technique.The results show that prestack fusion of seismic data can significantly enhance spatial sampling density and improve recognition capabilities for small fractured-vuggy bodies and small faults;after fusion,coverage increases by approximately 20%,signal-to-noise ratio increases by more than 10%,and recognition rate increases by 20%.Through selective fusion processing,near-offset data are applied to characterize small geologic bodies,while mid-and far-offset data support velocity analysis,thereby achieving effects comparable to high-density acquisition and full-offset imaging.This provides a degraded technical basis for secondary seismic acquisition without full-offset acquisition.This study provides a basis for optimizing observation-system design and guides secondary acquisition to obtain appropriate offset data,which can effectively reduce acquisition costs and has important practical significance for promoting efficient seismic exploration.

Key words: Tahe Oilfield, multiple acquisitions, fusion processing, high-coverage near offsets, efficient seismic exploration