Heating heavy oil reservoirs is a common method for reducing the high viscosity of heavy oil and thus increasing the recovery factor. Monitoring these changes in the reservoir is essential for delineating the heated region and controlling production. In this study, we measure the changes in the seismic wave attenuation of a heavy oil reservoir by constructing time-lapse Q −1 factor maps using a 4D-relative spectrum method. This method estimates seismic attenuation from surface reflection

more »

... surveys by calculating, for each trace in each survey, the attenuation (Q −1 ) using the spectral ratio (Toksoz et al. (1979) ) between a reference reflector above the reservoir and a second reflector below the reservoir. The results of our study on a real data set exhibit alignment along the injection wells, indicating that seismic attenuation can be used to monitor changes in a heavy oil reservoir. In recent years conventional crude oil reservoirs have been in decline and heavy oil is becoming an important potential resource. The production of conventional cold heavy oil at depths between 50 m and 1000 m has a typical recovery factor of 5% to 10% (Clark (2007) ). One method to increase recovery, is to heat a reservoir to above 200 • C either by combustion of part of the heavy oil (Vendati and Sen (2009), Kendall (2009) ) or by injecting steam into the reservoir (Clark (2007) ). This procedure makes heavy oil less viscous and more mobile. Theoretical and experimental studies show that the properties of heavy oil are strongly dependent on temperature, frequency, and composition (Eastwood (1993), Das and Batzle (2008) ). Behura et al. (2007) and Das and Batzle (2008) show that the shear modulus of heavy oil in general can be predicted by a frequency-dependent Cole-Cole visco-elastic model (Cole and Cole (1941) ), which has both real and imaginary attenuative parts. The frequency where the strongest attenuation is observed is called the relaxation frequency, and is related to the temperature through the viscosity of the oil (Das and Batzle (2008) ). From laboratory experiments, this peak attenuation is predicted to be at intermediate temperatures between 40 • to 120 • C. Because Batzle et al. (2006) indicates that heavy oils have different properties in the seismic, sonic, and ultrasonic frequency bands, these properties cannot be extrapolated from one band to another. Therefore, we need to have measurements in the seismic band in order to estimate attenuation for the intermediate temperatures. The measurement of seismic attenuation in the field is, in general, a difficult task because of the difficulty in discriminating the decay of the signal from attenuation and that from geometric spreading or reflection. The spectral ratio method, a common technique to estimate the attenuation (Q -factor) of the medium which separates the effect of attenuation from geometric spreading, was first presented for laboratory measurements by (Toksoz et al. (1979) ) and adjusted for vertical seismic profiles (VSP) by Hauge (1981) . The advent of time lapse surface seismic acquisitions using permanent systems with fixed positions for sources and receivers in heavy oil fields (Byerley et al. (2008) ), has made it possible to obtain high quality repeatable surface seismic data sets. Using such data we modify and adjust the standard spectral ratio method to the time-lapse surface reflection seismic data, and we show that changes in seismic attenuation due to the effect of steam injection can be monitored using this method. This paper is divided into two sections. In the first section, we present the 4D-Relative Spectrum Method (4DRSM). In the second section, we present results obtained with this method for a time-lapse reflection seismic data set from a heavy oil field in Athabasca, Canada.