The accuracy of estimations of wall concentration using the constant ( averaged) permeate velocity method was assessed in both empty and spacer-filled reverse osmosis (RO) channels under various operation conditions, based on the simulation results of a two-dimensional (2-D) concentration polarization model. It was determined that, although it always results in substantially larger variations of wall concentrations, the constant permeate velocity method is able to mimic the general variation patterns of wall concentrations in both empty and spacer-filled channels. The constant permeate velocity method can reasonably estimate the averaged wall concentration. The error is usually < 10%, under the simulation conditions. It was also determined that the constant permeate velocity method always overestimates the maximum wall concentration. For the membrane that is not in contact with transverse filaments, the error is usually within 20%, and increasing the mesh length or increasing the system value of Delta p/Delta pi(0) would generally decrease the error. For a membrane attached to ( touching) the transverse filaments, constant permeate velocity may erroneously overestimate the maximum wall concentration. The results suggest that, although the constant permeate velocity assumption may be acceptable for studying the overall wall concentration variation patterns or the averaged wall concentrations, this assumption may lead to erroneous results in the estimation of maximum wall concentration and, therefore, the possible onset of phase change of sparingly soluble salts in the concentration polarization layer.