Research paper Using sonic log data to investigate a multiple problem in the Hassi R'mel field, Algeria Zahia Benaïssa a, ⁎, Abdelkader Benaïssa a , Ferhat Salim Zaouia b , Tahar Aïfa c , Amar Boudella a , Nedjma Chami b , Djaffer Youcefi a , Mohamed Salim Ouadah b a Université des Sciences et de la Technologie Houari Boumediène, Faculté des Sciences de la Terre, BP 32, El-Alia, Bab-ezzouar, 16000 Alger, Algeria b Vsfusion, Baker Hughes-CGG Veritas Company, 20, rue Yahia Belhayat, Hydra, Alger, Algeria c Géosciences-Rennes, CNRS UMR6118, Université de Rennes1, Bat.15, Campus de Beaulieu, 35042, Rennes cedex, France abstract article info Article history: Received 11 April 2009 Accepted 10 November 2009 Keywords: VSP sonic log reflection primary multiple long-path multiple synthetics inside and outside corridor stack seismic section The conventional seismic reflection processing methods for detecting and suppressing multiples tend to fail in some circumstances. Vertical seismic profile (VSP) corridor stacks are now widely used to differentiate between primaries and multiples. However, they sometimes prove ineffective and the classical approach, using synthetics, becomes unoutlined. In the Hassi R'mel field, a strong event on the surface seismic sections, between clayey Dogger and lagoonal Dogger tops, is suspected to be a multiple because it is practically missing on the outside VSP corridor stacks. In this study, we show that the inside and outside VSP corridor stacks cannot help to identify this event. Using synthetics derived from well logs, acquired in six wells regularly distributed on the field, we conclude that the event is a primary. VSPs and sonic logs were recorded in the same well in every case. Synthetics, with multiples only, show an energetic long-path multiple with inverse polarity, at the same time as the event on the surface seismic sections. This explains: 1—the vanishing of the primary on the VSP corridor stacks, due to vertical recording in zero offset VSP acquisition, and 2—the efficiency of the standard common-midpoint (CMP) stacking to suppress the multiple on the surface seismic sections because of normal moveout differences between primaries and long-path multiples within the far-offset regions of data gathers. © 2009 Elsevier B.V. All rights reserved. 1. Introduction In seismic prospecting, the data quality is often degraded because of multiple reflections which make it difficult, and sometimes impossible, to interpret seismic sections. Much work has been done in developing methods for detecting and suppressing multiples (Mayne, 1962; Schneider et al., 1965; Michon et al., 1971; Taner, 1975; Lines, 1996; Voss and Hearn, 2003; Herrmann et al., 2007). These methods often face difficulties; common-midpoint (CMP) stacking methods based on normal moveout differences between multiple and primary reflections are ineffective when the offset range is small in comparison to the investigation depth. The predictive deconvolution method (Robinson and Treitel, 1980), which may work well for short-period multiples, is often unsuccessful for long-period multiples. Some filtering techniques are also ineffective, like f-k filter and τ-p filter (Yilmaz, 1987) if it is difficult to design accurate parameters. The use of well log data has proved to be effective in detecting insufficiently attenuated multiple reflections (Lines, 1996). Burton and Lines (1997) have examined a vertical seismic profiling (VSP) corridor stacking method for multiple identification, proposed by Hardage (2000), if reliable well log data are not available. This method, which therefore does not require well log data, is based on the differences between the inside and outside corridor stacks which can indicate the presence of short-period multiples. Hampson and Mewhort (1983) have applied a variation of this technique to long-period multiple attenuation. In the Hassi R'mel field, we suspect a strong arrival, appearing between clayey Dogger and lagoonal Dogger tops on the reflection seismic sections, to be a multiple because this event is without corresponding strong arrival on the outside corridor stacks. To verify it, we first used inside corridor stacks as proposed by Burton and Lines (1997) and then synthetics generated from well log data. In order to eliminate the influence of the well and to have data which satisfy our assumptions, six surveys, regularly distributed, with both VSP and sonic log recordings in every well, are considered here. Their approximate locations are given by numbers 1 through 6 in Fig. 1. Fig. 2 shows the tie between the VSP corridor stack of well 1 and the seismic record section passing through this well. The suspected event, marked by the box in the section depicted in Fig. 2, appears at 1010 ms. 2. Acquisition of field data VSP and well log data discussed here were acquired by the Sonatrach Company in the six wells of the Hassi R'mel field (Fig. 1). Zero offset VSP surveys were carried out over depth ranges varying from 1240 to 1369 m, starting at depths from about 895 to 1000 m. The spacing of the recordings was held constant at intervals of 20 m for wells 1, 2, 3 and 6 Journal of Petroleum Science and Engineering 71 (2010) 121–125 ⁎ Corresponding author. E-mail addresses: zabendz@hotmail.fr (Z. Benaïssa), tahar.aifa@univ-rennes1.fr (T. Aïfa). 0920-4105/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.petrol.2009.11.008 Contents lists available at ScienceDirect Journal of Petroleum Science and Engineering journal homepage: www.elsevier.com/locate/petrol