The South Caspian Basin (SCB) is known as one of the World\\\'s deepest and largest hydrocarbon-prone basins. Due to high subsidence and sedimentation rate in this deep-water basin, many overpressure sandy zones are susceptible to shallow water flow (SWF) problem. SWF is one of the geological hazards in deep-water exploratory drilling such as SCB. When the drill reaches the SWF zones, sands and water flow into the wellbore which cause the drilling failure. Therefore, the present study is the first detailed evaluation of SWF problem predrilling in the SCB, offshore Iran by using 3D post-stack, pre-stack seismic data and log data. Seismic attributes have been applied to identify the buried channels which may be considered as the most potential feature for SWF occurrence. Simultaneous inversion has been used to extract the parameters such as P-impedance (AI), S-impedance (SI), Vp/Vs ratio, Poisson\\\'s ratio. Based on inversion results, unsupervised artificial neural networks also used to estimate the probability of SWF risk in the buried channel for the first time. Our initial evaluations based on the similarity and sweetness attributes indicated that some buried shallow sandy channels with a variable width between 500 and 1500 m covered by imperious clay layers are the most potential facies for creating the SWF zones in the SCB. The SWF zone is detected by low-impedances (AI<3150 (m/s)*(g/cc), SI < 1500(m/s)*(g/cc), high Vp/Vs ratio (1.21 to 2.23) and Poisson\\\'s ratio in range of 0.05 to 0.38. The probability of SWF risk of the detected channel is estimated to be >0.35 by unsupervised artificial neural networks. Finally, our findings provide a new comprehensive evaluation of SWF problem in the channel system predrilling in the SCB which may be useful for other researchers and decrease the drilling risk in future works.

Keywords