Available online at www.sciencedirect.com Physica A 318 (2003) 551–561 www.elsevier.com/locate/physa Seismic ground roll time–frequency ltering using the gaussian wavelet transform G. Corso ∗ , P.S. Kuhn, L.S. Lucena, Z.D. Thom e Departamento de F sica Te orica e Experimental, International Center for Complex Systems, Universidade Federal do Rio Grande do Norte, Natal CEP 59078-970, Brazil Received 7 December 2001 Abstract Seismic signal processing is an important task in geophysics sounding and represents a per- manent challenge in petroleum exploration. Although seismograms could in principle give us a picture of a geological structure, they are very contaminated by spurious signals (having the ground roll as the main component). This fact demands a big eort in developing new ltering methodologies. Using the Gaussian wavelet transform, a ltering method for ground roll removal is developed. The lter allows a local extraction of the ground roll, it is adaptative to trace and it has an attenuation factor that keeps the average frequency spectrum. This method is tested for a land-based seismic signal leading to promising results. c 2002 Elsevier Science B.V. All rights reserved. PACS: 43.50.+y; 91.30.Dk; 43.60.+d; 93.85.+q Keywords: Seismic processing; Noise ltering; Wavelets; Ground roll; Petroleum exploration 1. Introduction Petroleum reservoirs are complex systems whose analysis is well tted for the meth- ods of statistical physics. One of the main problems of the industry is to nd petroleum occurrences that are economically feasible. This is the goal of petroleum exploration. The experimental data are obtained directly from exploration wells or through seismic wave soundings. A big challenge for scientists is the cleaning of seismic data which are contaminated by noise, mainly the ground roll [1–5]. Ground roll is a common * Corresponding author. E-mail addresses: corso@dfte.ufrn.br (G. Corso), kuhn@dfte.ufrn.br (P.S. Kuhn), liacir@dfte.ufrn.br (L.S. Lucena), zielithome@ig.com.br (Z.D. Thom e). 0378-4371/03/$-see front matter c 2002 Elsevier Science B.V. All rights reserved. PII:S0378-4371(02)01379-1