Passive seismic low-frequency ͑from approximately 1-6 Hz͒ data have been acquired at several locations around the world. Spectra calculated from these data, acquired over fields with known hydrocarbon accumulations, show common spectral anomalies. Verification of whether these anomalies are common to only a few, many, or all hydrocarbon reservoirs can be provided only if more and detailed results are reported. An extensive survey was carried out above a tight gas reservoir and an adjacent exploration area in Mexico. Data from several hundred stations with three-component broadband seismometers distributed over approximately 200 km 2 were used for the analysis. Several hydrocarbon reservoir-related microtremor attributes were calculated, and mapped attributes were compared with known gas intervals, with good agreement. Wells drilled after the survey confirm a predicted high hydrocarbon potential in the exploration area. A preliminary model was developed to explain the source mechanism of those microtremors. Poroelastic effects caused by wave-induced fluid flow and oscillations of different fluid phases are significant processes in the low-frequency range that can modify the omnipresent seismic background spectrum. These processes only occur in partially saturated rocks. We assume that hydrocarbon reservoirs are partially saturated, whereas the surrounding rocks are fully saturated. Our real data observations are consistent with this conceptual model.

A B S T R A C TResults of a passive microtremor survey at an oil and gas field in Voitsdorf, Austria, are presented. The survey consists in six parallel profiles approximately 9 km long over two hydrocarbon reservoirs. For each profile the seismic wavefield was recorded synchronously at 11 in-line stations. The measurements were conducted with broadband seismometers and lasted, for each profile, at least 12 hours overnight. Data interpretation is based on a comprehensive data set and on the analysis of four different spectral attributes. These attributes quantify the characteristic features of the wavefield's Fourier spectra in the low-frequency range (<10Hz). One attribute quantifies the spectral energy in the vertical wavefield component, another attribute quantifies the maxima in vertical-to-horizontal spectral ratios and two attributes describe the frequency shifts of peaks within the spectra of vertical and horizontal wavefield components. Due to temporal variations of the signals we combine the long-term measurements (several hours of continuous records) of multiple profiles. This procedure considerably enhances the consistency of each spectral attribute and makes them suitable to quantify lateral variations of the wavefield. The results show that using a combination of several attributes significantly increases the reliability of detecting anomalies in the microtremor wavefield that are presumably caused by hydrocarbon reservoirs. A numerical study of two-dimensional seismic wave propagation is applied to investigate the peak frequency shift attributes. The results of the study indicate that the attributes may contain information on the depth of hydrocarbon reservoirs, assuming that the reservoir acts as a (secondary) source of low-frequency seismic waves.

We welcome the opportunity to respond to Professors Green and Greenhalgh's analysis of our work in this journal. In addressing their observations we hope to contribute towards a better understanding of our work in general.Green and Greenhalgh (2009) commented not only on Lambert et al. (2009) but also on peer-reviewed papers, conference proceedings and magazine articles authored by at least one of us, thereby combining formative early work with more advanced results in their critique. Additionally, they commented on publications that none of us has co-authored (e.g., Dangel et al. 2003;Walker 2008), addressing questions to a scientific community beyond our research group. This reply necessarily focuses only on the work that at least one of us has co-authored.Green and Greenhalgh (2009) asserted that our conclusion that "the results indicate that passive low-frequency spectral analysis can increase the probability of locating reservoirs significantly" is unfounded. We identify three main arguments in their commentary: i) that our attributes are not applicable to reservoir detection, ii) that the results of our numerical study do not fit the data and iii) that our attributes are in fact sensitive to shallow structures and we inadequately consider these potential shallow effects in our work. We first respond to these three main points and then to additional criticism, which is included as discussion within their commentary. Finally, we describe the objectives of our research to put it in the correct perspective and to address Green and Greenhalgh's (2009) as- * Green and Greenhalgh (2009) expressed the view that our Attributes 1 and 2 do not indicate the two reservoir locations. Based on our qualitative analysis of the correlation between these spectral attributes and reservoir locations, we agree with them. In Lambert et al. (2009) we explicitly stated: "Attributes 1 and 2 are not sensitive for the northern reservoir", "For Attribute 1, temporal variations are rather large and, therefore, the observed anomalies are less significant" and "Attribute 2 behaves very stable in time . . . the profiles do not show clear lateral anomalies". We clearly stated that the profiles of Attributes 1 and 2 are not useful for detecting the two reservoirs at this site. We do not see the supposed contradiction in Lambert et al. (2009) because we did not claim that all attributes must show anomalies above both reservoirs in order to increase the probability of detecting reservoirs. Attributes 3 and 4Attributes 3 and 4 are based on more stable frequency values rather than on amplitude values, which can vary substantially in experiments utilizing an uncontrolled source. Green and Greenhalgh (2009) wrote that "Lambert et al.'s (2009) simplistic physical exploration is both implausible and inconsistent with (our) observations". Our numerical study was in fact a feasibility (or plausibility) study intended to show that seismic signals emitted by a subsurface source in a homogeneous, C 2009 European