|dc.description.abstract||In this study, I utilize multicomponent time-lapse seismic datasets for investigating subtle seismic properties of Weyburn reservoir undergoing enhanced oil recovery and geologic sequestration of CO2. The primary focus is on extracting shear-wave information from surface three-dimensional and three-component (3-D/3-C) reflection datasets. Four groups of interrelated objectives are addressed: 1) calibrated and true-amplitude processing of multicomponent time-lapse seismic data, 2) extraction of amplitude variations with angle (AVA) and offset (AVO) attributes for separating pressure and fluid-saturation effects within the reservoir, 3) development of receiver-function methods for investigating the shallow subsurface, and 4) 2-D spatial pattern analysis of attribute maps, intended for automated interpretation of the results and a new type of AVO analysis.
To achieve the first of these objectives, I reprocess the field surface 3-C/3-D reflection datasets by using pre-stack waveform calibration followed by complete reflection processing using commercial ProMAX software. For the second, principal objective of this study, several AVA attributes of the reservoir are examined, including those related to P- and P/S- converted waves and P- and S-wave impedances. The amplitudes and AVA attributes derived from seismic data indicate temporal variations potentially caused by pore-pressure and CO2-saturation variations within the reservoir. By comparing with AVA forward models, the seismic data suggest correlations between the increasing pore pressure and decreasing AVA intercepts and increasing AVA gradients. Increasing CO2 saturations appear to correlate with simultaneously decreasing AVA intercepts and gradients. CO2-saturated zones are thus interpreted as Class III AVA anomalies.
In order to take further advantage from 3-C recordings and investigate advanced methods for S-wave seismic data analysis, receiver functions are used to study the shallow near-surface structure. This is apparently the first application of this method to reflection seismic datasets on land and in a time-lapse 3-D dataset. I show that it is feasible and useful to measure the near-surface S-wave velocity structure by using multi-component seismic data. From Weyburn reflection data, the average mapped receiver-function time lags are about 35 ms, which corresponds to near-surface S-wave velocities of about 550 m/s. Time-lapse variations of the near-surface structure are measured, and S-wave statics models are derived. Such models can be useful for converted-wave seismic imaging.
The last objective of this Dissertation is to develop tools for interpretation of gridded 2-D spatial images, such as mapping AVO attribute quantitatively and automatically. For this purpose, a new pattern-recognition approach called skeletonization is developed and applied to several regional aeromagnetic and gravity images from southern Saskatchewan and Manitoba. The approach is combined with 2-D empirical mode decomposition allowing pattern analysis at variable spatial scales. The results show that skeletonization helps identifying complex geologic structures and measuring their quantitative attributes that are not available from conventional interpretation. Applications of this approach to interpretation of AVO attributes are discussed.||