Mortazavi, S.A., Javaherian, A., Bidhendi, M.N., Amindavar, H.R., Torabi, S$. and Bakhtiari, M.R., 2017. Application of adaptive pyramidal dual-tree directional filter bank in ground roll attenuation. Journal of Seismic Exploration, 26: 49-79. Ground roll is a coherent noise that may mask reflections in a fan-shaped zone. Its attenuation, accordingly, is of utmost significance in seismic data processing. Different methods have been developed to attenuate ground roll. In this paper, an adaptive pyramidal dual-tree directional filter bank (PDTDFB) was applied to a synthetic data consisting of ground roll, aliased ground roll, and reflections over an earth model and implemented on two shot records from the west and south of Iran. PDTDFB is a multiscale and a multidirectional filter with a dual-tree structure which decomposes an image (or a shot record) to various directional subscales in different scales. Certain subscales predominantly contain ground roll, and aliased ground roll, having higher values of energy than others. In the first step, through adaptively detecting these subscales via energy calculation, ground roll and aliased ground roll are attenuated. In the second step, the remaining aliased ground roll is adaptively attenuated. According to the results of synthetic and real data, the adaptive PDTDFB highly attenuated the ground roll and aliased ground roll with minimum harm to signals. We further examined the effects of random noise and residual statics when using the adaptive PDTDFB in ground roll attenuation. The adaptive PDTDFB is not too sensitive to the level of random noise. Moreover, the performance of the filter is acceptable in the presence of residual statics. Also, according to the comparison with f-k and SVD filters and the qualitative and quantitative assessments, the proposed filter entailed better results than f-k and SVD filter, especially as far as aliased ground roll attenuation is concerned.