Abstract

The Mumbai High Field, one of India’s largest and mature offshore oil fields, consists of complex, multi-layered reservoirs. The application of spectral decomposition in this context holds significant potential for improving reservoir characterization. It facilitates the delineation of thin beds, and stratigraphic discontinuities, thereby providing deeper insights into the subsurface architecture and aiding in more informed exploration and development strategies. This paper presents a comprehensive comparative analysis of three fundamental spectral decomposition methods employed in seismic data: Short-Time Fourier Transform (STFT), Continuous Wavelet Transform (CWT), and Matching Pursuit (MP). Each method is examined through rigorous mathematical foundations, computational complexity analysis, and practical applications in subsurface characterization. STFT provides efficient time frequency analysis with fixed resolution constraints, while CWT offers adaptive multiscale decomposition capabilities. MP techniques deliver sparse signal representations with superior interpretability and compression ratios. Our analysis demonstrates that optimal seismic processing requires intelligent integration of these complementary approaches, with method selection dependent on specific geological objectives and computational constraints. Performance evaluation using synthetic and real seismic data validates the theoretical predictions and provides practical implementation guidelines for the Geophysical community. In the current study, a prominent channel feature has been delineated at the Late Oligocene (L-I) reservoir level. The study shows various time slices and figures for improvement in well positioning.