A Novel Workflow to Invert Broadband Seismic Data: A Case Study from Onshore Fields in Abu Dhabi

Abstract

Abstract Broadband seismic data has several benefits for quantitative seismic reservoir characterization. It is characterized by a significant increase of seismic frequency bandwidth on both the low and high sides of the frequency spectra. This work presents a novel seismic inversion approach to exploit the full value out of broadband seismic data. The average wavelet from broadband seismic data is limited in high and low frequencies due to the short duration of the well log and the misalignment of the seismic data with the well-log synthetic at high frequencies. Limitation of the extracted wavelet and optimization can generate band-limited inversion results that do not capture the full range of frequencies. An alternate approach of dividing the data into three frequency bands resulted in extracted wavelets that capture the spectrum of each band, and in turn produced a reliable broadband inversion result honoring the full range of frequencies present in the data. Inversion results gave a superior match of the estimated synthetic with the data spectra (Figure 1), and the reservoir was better calibrated at all the well locations. Successful recovery of the ultra-low frequencies enabled us to maximize the value of the broadband data. The workflow also pushed the frequency of the inverted properties to 80 Hz which helped in turn to characterize some of the relatively thinner layers, which were otherwise getting averaged out. Building a low frequency model for AVO seismic inversion using ultra-low frequency information leads to a significant improvement of predictability away from wells. As a prior model, a geologically constrained (4 Hz) low frequency filter was applied. Review of the broadband AVO seismic inversion results clearly indicate a better match between the inverted traces and well log properties at the studied wells. Also, the blind well test results at four wells indicate an excellent match to the blind well logs, which adds a high degree of confidence on the inverted elastic properties. Also, the synthetic spectra of the ultra-low and ultra-high frequencies is captured and maintained in the inverted broadband seismic data. The novelty of the new workflow is in the ability to effectively invert the broad frequency band of seismic data. Successful recovery of the ultra-low and ultra-high frequencies enabled us to maximize the value of the broadband data. Subsequently, the high frequency elastic properties helped in successful characterization of thinner reservoirs and will help in better optimization of the future field development initiatives.