It has long been recognized that repeated gravity surveys could be used to track changes in either elevation or mass distribution in the Earth. The technology to effectively track mass distribution changes resulting in μGal level signals over long periods of time has matured in recent years. The Prudhoe Bay, Alaska, water flood surveillance project has hastened that development and set a new standard for the conduct of time lapse or 4-D gravity surveys. This talk will review the history of the Prudhoe Bay efforts and some of the milestones achieved along the way.

The Prudhoe Bay reservoir water flood is the largest ever undertaken and is intended to repressurize the gas cap and maintain declining production over a period of decades. It is difficult to monitor the progress of the water due to a lack of wellbores located in the gas cap. The 4-D seismic method will also be used, but it is limited by expense and permafrost. In 1993 Jerry Brady and Don Walcott, then at ARCO Alaska, started to consider the application of repeated surface and borehole microgravity surveys to monitor the water flood.

The University of Texas at Dallas (UTD) became involved in a theoretical investigation of the possibility of gravity surveillance. An inversion procedure was formulated and tested on synthetic gravity data based on reservoir simulations. Various 4-D gravity noise scenarios were proposed and the resolution of the method determined. At about the same time (1994), a program of field experiments was initiated to refine procedures for actually obtaining the type of data required for the modeling. It soon became clear that the state of the art would require some extension to achieve that goal.

In successive field experiments, conducted in the Arctic winter, microgravity measurement techniques (both relative and absolute gravity meters) and geodetic measurements using the Global Positioning System were refined. The noise levels to be expected in the 4-D gravity data were characterized and a long-term monitoring program was planned, involving about 300 stations. In 2002 a full-scale baseline survey was conducted and late in that year water injection commenced. Repeat surveys were conducted in 2003 and 2005, and a third survey is planned for this year.

The 4-D data over the 2002 to 2005 interval has been modeled and the water flood has been detected. Model results resemble predictions from reservoir simulations but are also producing unexpected results that should help the reservoir engineers understand the actual situation in the ground. The methodologies and standards developed for this project are now being used to plan surveys in other areas.

End_of_Record - Last_Page 21---------------