Estimation of free gas saturation from seismic reflection surveys by the genetic algorithm inversion of a P-wave attenuation model

Many previously proposed methods of estimating free gas saturation from seismic survey data rely on calibration to invasively collected, in situ measurements. Typically, such in situ measurements are used to parameterize or calibrate rock-physics models, which can then be applied to seismic data to achieve saturation estimates. We tested a technique for achieving estimates of the spatial distribution of gas saturation solely from shipboard seismic surveys. We estimated the quality factor from seismic reflection surveys using the spectral ratio method, and then inverted a mesoscopic-scale P-wave attenuation model to find the parameters that matched the modeled attenuation to our estimates of observed attenuation within the range of seismic frequencies. By using a genetic algorithm for this inversion, we not only searched efficiently for a global solution to the nonlinear set of equations that compose the model, but also constrain the search to a relatively broad set of realistic parameter values. Thus, our estimates do not rely on in situ measurements of these parameters, but on distributions of their possible values, many of which may be referenced from literature. We first tested this method at Blake Ridge, offshore North and South Carolina, where an approximately 400-m-deep gas-saturated zone underlies a field of methane hydrates. The extensive field work and subsequent studies at this site make it ideal for validating our method. We also demonstrated the applicability of our method to shallower deposits by presenting results from Finneidfjord, Norway, where the inversion of the P-wave attenuation model recognizes very small gas saturations.