Quantitative seismic analysis of a thin layer of CO2 in the Sleipner injection plume

Time-lapse seismic reflection data have proved to be the key monitoring tool at the Sleipner [Formula: see text] injection project. Thin layers of [Formula: see text] in the Sleipner injection plume show striking reflectivity on the time-lapse data, but the derivation of accurate layer properties, such as thickness and velocity, remains very challenging. This is because the rock physics properties are not well-constrained nor are [Formula: see text] distributions on a small scale. However, because the reflectivity is dominantly composed of interference wavelets from thin-layer tuning, the amplitude and frequency content of the wavelets can be diagnostic of their temporal thickness. A spectral decomposition algorithm based on the smoothed pseudo Wigner-Ville distribution has been developed. This enables single frequency slices to be extracted with sufficient frequency and temporal resolution to provide diagnostic spectral information on individual [Formula: see text] layers. The topmost layer of [Formula: see text] in the plume is particularly suitable for this type of analysis because it is not affected by attenuation from overlying [Formula: see text] layers and because there are areas in which it is temporally isolated from deeper layers. Initial application of the algorithm to the topmost layer shows strong evidence of thin-layer tuning effects. Analysis of tuning frequencies on high-resolution 2D data suggests that layer two-way temporal thicknesses in the range 6 to 11 ms can be derived with an accuracy of c. 2 ms. Direct measurements of reflectivity from the top and the base of the layer permit calculation of layer velocity, with values of around [Formula: see text], in reasonable agreement with existing rock physics estimates. The frequency analysis can, therefore, provide diagnostic information on layer thicknesses in the range of 4 to 8 ms. The method is currently being extended to the full 3D time-lapse data sets at Sleipner.