Abstract: Based on wavelet theory, this paper obtained the multi-scale dominant frequency (MSF)by using wavelet analysis, and derived a new relationship algorithm between travel-time, frequency and absorption coefficients from transforming the seismic amplitude equation proposed by Aki & Richards (1980), then calculating the multi-scale absorption attributes (MSA)of seismic signals in time-frequency domain, which effectively show the oil and gas response characters, such as low main frequency, serious high frequency absence, dimming reflection, reflection blank, reflection disorder and fat wave shape. Hydrocarbon reservoir has special response to high frequency sects, and goes with intensive amplitude, energy and frequency absorption of seismic waves. Calculating MSF and MSA of seismic data at different scale can reveal frequency shift and absorption attenuation characters of seismic waves, which is favorable to lithology identification, petroleum reservoir prediction and synthetic interpretation. A case study of MSF and MSA techniques applied to seismic data from dense clastic rocks of Xujiahe formation in depression of western Sichuan basin provides robust indications for reservoir prediction and gas detection.

Key words: wavelet transform, seismic signal, frequency, absorption attributes, hydrocarbon reservoir