Land seismic repeatability prediction from near-surface investigations at Naylor Field, Otway

Abstract

Time-lapse seismic is a powerful methodology for remotely monitoring changes in oil and gas reservoirs. Its high sensitivity and resolving power make it the methodology of choice for monitoring CO2 sequestration in deep saline aquifers or depleted oil and gas fields. This method is now routinely applied offshore but rarely onshore because of inherently poor repeatability of land seismic data. Considering that CO2 sequestration on land is becoming a necessity, there is a great need to evaluate the feasibility of this method for land based CO2 sequestration projects. A feasibility study, onshore Otway Basin, Australia, aims at evaluating the viability of monitoring methodologies for the case of CO2 storage into a depleted gas field. Since injection of CO2 into a depleted gas field at a depth of around 2 km causes very subtle changes in elastic properties of the reservoir rock, it is critical to achieve high repeatability of time-lapse seismic surveys if they are to be implemented into a monitoring program.The goal of this thesis is to analyse the main factors affecting seismic repeatability at the Otway site. I aim to achieve this goal through the deployment of pre-base line measurements and combining the results with detailed numerical modelling studies. Such measurements have to be rapid, effective and quantitative so that a seismic monitoring team can decide whether to use time-lapse methodology when processing their data.To find the most likely repeatability at the Otway site I used so-called micro-arrays (surface and borehole) in a time-lapse manner to determine the seasonal variation of elastic properties of the near surface. The measurements were aimed at determining directional P-wave velocity and attenuation (Q-factor). The top soil (0.5m thick agricultural layer or elasto-plastic zone) had a low velocity and low Qfactor and hence significantly attenuated seismic energy.The elastic parameters obtained were then used to numerically simulate real timelapse surveys. The results obtained were compared and verified against conventional time-lapse studies conducted at the Otway site over a three year period, at different times of the year and with different sources. The agreement between numerical and field data, expressed through a normalised root mean square (NRMS) difference confirms that the effect of the near surface variation in the time-lapse land seismic can be predicted with minimum cost and through the deployment of small, inexpensive experiments.