Achieving uniform mixing conditions are essential for the fioccuiation process to optimize floe size and avoid floc-breakup. Limited literature is available on establishing consistent operational conditions and procedures for pilot-scale fioccuiation systems, which have tank sizes smaller than full-scale and larger than jar-test equipment, in this study, the influence of mixing speeds on the determination of the conventional design parameter, the average velocity gradient (G), was investigated for pilot-scale paddle flocculators. The pilot-scale plant for this paper was hosted at the J.D. Kline water Supply Plant (JDKWSP) in Halifax, Canada. Computational fluid dynamics (CFD) was evaluated as an alternative design technique and compared against traditionally used empirical-based calculations. Comparison of both approaches showed that the G-values of empirical method were substantially higher than the predicted values for rotational speeds greater than 5 rpm. In contrast, CFD predictions found that G-values used for tapered paddle fioccuiation process (up to 60s"^) could be achieved at lower rotational speed (around 15 rpm), which minimizes the power input required for mixing. The practical implications of operating at higher than required G-values relates to potential negative consequences such as floe break up, and the reliance of chemical additives to avoid floe break-up. These very practical outcomes could impact the interpretation of findings from pilot-scale treatment systems. NOMENCLATURE Ap C CD CFD FE FTC FV FD G projected area of paddle (m^) mean tracer concentration (kgm^) drag coefficient on paddle for turbulent flow (dimensionless number) computational fluid dynamics finite element flow through curve finite volume drag force on paddles global root mean square velocity gradient or energy input rate (s~^)