Hamidi, R., Javaherian, A. and Reza, A.M., 2013. Comparison of IDWT and 2DWT transforms in ground roll attenuation. Journal of Seismic Exploration, 22: 49-76. Ground roll is the most important type of coherent noise in land seismic data. It usually has stronger amplitude than reflections and masks the valuable information carried by the signals. Wavelet transform is one of the methods which can be used for ground roll suppression. It can be used in one or two dimensions based on the nature of data which is supposed to be filtered. In one-dimensional wavelet transform (1DWT) any 1D trace is transformed into a time-scale domain. This enables the separation of features with different frequencies in different coefficients while preserving their time separation. Two-dimensional wavelet transform (2DWT) in fact takes two 1DWTs in columns and rows of a data matrix. In this case, seismic events with different velocities are represented in different horizontal, vertical and diagonal detail coefficients. It is enough to determine the coefficients corresponding to the noise and omit them to have the filtered data. In this study, 1DWT and 2DWT were both reviewed, the MATLAB code for two filters was written to suppress the ground roll, and the results of filtering the synthetic and real data are presented. The synthetic data was based on an earth model of eight layers over a half space containing refraction, reflectors and a ground roll. The real data is a shot record from SW Iran with a strong ground roll. According to the results, 2DWT can extract the ground roll with less suppression of the desired signals compared to 1DWT. It is due to the fact that the noise separation from signals has a better resolution in the 2D case. The filters were based on the frequency content (in IDWT), arrival time and velocity (in 2DWT) of the ground roll and signal. For that reason, these were the filter parameters (which depended on the input data). To have an appropriate result, at least one of these properties had to be different. In addition, the choice of the mother wavelet could affect the filter performance because different wavelets produced different results in the separation of signal and noise ina WT domain.