Contemporary three-dimensional numerical sediment transport models are often computationally expensive because of their complexity and thus a compromise must be struck between accurately modeling sediment transport and the number of effective sediment grain (particle) size classes to represent in such a model. The Environmental Fluid Dynamics Code (EFDC) was used to simulate the experimental results of previous researchers who investigated sediment erosion and gradation around a 180° bend subject to transient flow. The EFDC model was first calibrated using the eight distinct particle size classes reported in the physical experiment to find the best erosion formulations to use. Once the best erosion formulations and parameters were ascertained, numerical simulations were carried out for each experimental run using a single effective particle size. Four techniques for evaluating the effective particle size were investigated. Each procedure yields comparable effective particle sizes within a factor of 1.5 of the others. Model results indicate that particle size as determined by the weighted critical shear velocity most faithfully reproduced the experimental results for erosion and deposition depths. Subsequently, model runs were conducted with different numbers of effective particle size classes to determine the optimal number that yields an accurate estimate for noncohesive sediment transport. Optimal, herein, means that numerical model results are reasonably representative of the experimental data with the fewest effective particle size classes used, thereby maximizing computational efficiency. Although modeling with more size classes can be equally accurate, results from this study indicate that using three effective particle size classes to estimate the distribution of sediment sizes is optimum.