Continuous monitoring by ambient-seismic noise tomography

Introduction

Reservoir monitoring by active seismic is a well-established technique. A single survey provides an image of the subsurface at a given time. Repetitive seismic surveying provides differential seismic images that show how the subsurface changes over time, relative to a baseline survey. At Valhall, in the Norwegian North Sea, a permanent four component (4C) Ocean Bottom Cable (OBC) array was installed in 2003 (Kommedal et al., 2004) for repeated active seismic surveying. These conventional surveys provide subsurface images after lengthy seismic acquisition, processing, and imaging procedures, at only a few snapshots in time. The OBC array is capable of continuously recording for long periods of time in the presence or absence of active seismic shooting.

Recordings in the absence of controlled-seismic shooting are referred to as passive-seismic recordings. Their use is widespread in both global- and exploration-seismology. Passive-seismic recordings contain earthquake energy of rupture events as large as the 2004 Sumatra Earthquake, $M_w=9.1$ , (Ammon et al., 2005), and as small as hydraulic induced microseismic fracture events in reservoirs, $M=-2$ , (Shemeta and Anderson, 2010). Studies of earthquakes and microseismic events are of deterministic nature, aimed at characterizing the source and its wavefield. However, even in the absence of discrete deterministic events, passive-seismic recordings contain an abundance of energy, also referred to as the Earth's Hum (Rhie and Romanowicz, 2004,2006). These are continuous excitations of the Earth's free modes. One strong such series of excitations is the microseism energy captured in the Earth's surface interface modes. The waves in these modes are propagating randomly, as such no particular events can be distinguished. Hereafter this microseism energy will also be referred to as ambient-seismic noise.

Recordings of microseism energy can be made continuously and thus are a potential source of data for continuous reservoir monitoring. This is not only of interest for monitoring production-related changes over a long time scale, but also a possible source of data for short time scale hazard monitoring.

Passive seismic interferometry is a technique that cross-correlates ambient-seismic noise recordings at two stations to form a signal, the estimated Green's function (EGF), as if one of the stations was a seismic source (Wapenaar and Fokkema, 2006). Virtual seismic Scholte wave sources have previously been used in tomographic studies of the near surface (Bussat and Kugler, 2009; de Ridder and Dellinger, 2011). This technique allows for the design and installation of a system that exploits the ambient-seismic noise field to perform continuous, nearly real-time reservoir surveillance. The reliability of retrieved subsurface information depends on the characteristics of the ambient seismic field and its excitation sources. We need to investigate how much ambient-seismic noise is needed to cross-correlate for the retrieve of stable EGFs and reliable tomographic images.

This paper is outlined as follows. In the first section we briefly review passive seismic interferometry. Then we explore the ambient seismic microseism energy at Valhall and its capability to create low frequency virtual seismic sources. In the following section we elaborate on the filtering and correlation procedures. We then explore with how much data the correlations convergence towards the long-term correlation average. The next section contains straigth-ray tomography results at various frequency bands and for correlations of various portions of the data. The paper finishes with a discussion and some concluding remarks.

Continuous monitoring by ambient-seismic noise tomography