The seismic reflection method is a well-established method for imaging selected parts of the Earth’s subsurface with high resolution and high accuracy. In this method man-induced sources are used. The registrations are processed by advanced seismic imaging algorithms and yield a 2- or 3-dimensional image of the subsurface. In the exploration for hydrocarbons the image is usually 3-dimensional, the maximum depth of the image is 4 to 5 kilometers, the horizontal dimensions are typically 20 by 20 kilometers and the vertical resolution is in the order of 40 meters. On the other hand, for the analysis of the crystalline crust the seismic acquisition is usually done along one or more lines, yielding 2-dimensional image(s) with a maximum depth in the order of 30 kilometers, a length of several hundreds of kilometers and a vertical resolution in the order of 300 meters. Current seismic reflection methods are not suited for imaging large 3-dimensional volumes like, for example, the subsurface of the Netherlands up to a depth of 30 km. The limitation of the image volume is not a matter of principle but of logistics and costs. This even more applies to seismic monitoring of processes in the subsurface, which requires a repetition of the seismic acquisition and imaging procedure from time to time. This so-called time-lapse seismics is nowadays applied in seismic exploration, but is unthinkable at the large scale we just mentioned.
A revolutionary passive seismic method will be used which employs natural sources (transient signals from small earthquakes as well as natural noise signals) in the subsurface and thus circumvents the extensive use of man-induced sources. The passive seismic method we refer to aims at retrieving seismic reflection measurements from passive recordings at the Earth’s surface of the transmission response of natural sources in the subsurface. This idea was first proposed by Claerbout in 1968, for horizontally layered media. Later Claerbout conjectured for arbitrarily inhomogeneous media that “by cross-correlating noise traces recorded on two locations on the surface, we can retrieve the wavefield that would be recorded at one of the locations if there was a source at the other”. The conjecture was first proved in 2003 (Wapenaar 2003). The proof is illustrated for the subsurface structures under Annerveen ( Figure ) using numerical modelling. A number of uncorrelated noise sources (the yellow stars) in the subsurface radiate seismic waves to the surface, where they are recorded by seismic sensors (the yellow triangles) at different locations. The result of such a recording is shown in the left part of Figure 4 and represents the a few seconds of a 2-hours-long transmission noise recording, The causal part of the cross-correlation of two noise traces at, e.g., 17 km and 10 km yields the reflection response observed at the receiver at 10 km as if there was an impulsive source at the receiver at 17 km. When there are sufficient receivers measuring the natural signals, like in our example from 12 km until 22 km every 50 m, one could in principle retrieve the seismic reflection response for many source and many receiver positions and subsequently obtain the structures in the subsurface as is normally done with active seismic measurements. In the right part of we show the retrieved reflection response for a virtual source at 17 km at the surface and receivers from 12 km until 22 km.
Figure 3. Model for subsurface near Annerveen, with receivers at surface (triangles) and simulated natural sources (stars)
Figure 4. Transmission responses from sources (left) and retrieved reflection response (right), for model near Annerveen
Until recently, the application of seismic interferometry to passive seismic recordings was limited to retrieval of surface waves. Surface waves, generated by tidal waves, have been used to construct direct surface-wave seismograms and subsequently inverted for subsurface structure. However, the resolution of the subsurface image that can be obtained using surface waves is inferior to the resolution from reflected waves. Only recently, Draganov et al. (2009) showed that one could retrieve the Earth’s reflection response from recorded seismic noise. They showed these using passive recordings from Libya. The retrieved reflections were then used to estimate the subsurface velocity structure and finally to obtain a 3-D image of subsurface reflectors.
These first results encourage us to proceed with the phased program of LOFAR via simulation, prototyping and demonstration, with the ultimate goal of developing a unique permanently installed, densely instrumented, passive seismic imaging array. The goals are: