The application of acoustic Doppler current profilers (ADCP's) in river flow measurements is promoting a great deal of progress in hydrometry. ADCP's not only require shorter times to collect data than traditional current meters, but also allow streamflow measurements at sites where the use of conventional meters is either very expensive, unsafe, or simply not possible. Moreover, ADCP's seem to offer a means for collecting flow data with spatial and temporal resolutions that cannot be achieved with traditional current-meters. High-resolution data is essential to characterize the mean flow and turbulence structure of streams, which can in turn lead to a better understanding of the hydrodynamic and transport processes in rivers. However, to properly characterize the mean flow and turbulence intensities of stationary flows in natural turbulent boundary layers, velocities need to be sampled over a long-enough time span. The question then arises, how long should velocities be sampled in the flow field to achieve an adequate temporal resolution? Theoretically, since velocities cannot be sampled over an infinitely long time interval, the error due to finite integration time must be considered. This error can be estimated using the integral time scale. The integral time scale is not only a measure of the time interval over which a fluctuating function is correlated with itself but also a measure of the time span over which the function is dependent on itself. This time scale, however, is not a constant but varies spatially in the flow field. In this paper we present an analysis of the effect of the temporal resolution (sampling time span) on the accuracy of ADCP measurements based on the integral time scale. Single ping velocity profiles collected with frequencies of 1 Hz in the Chicago River at Columbus Drive using an uplooking 600 kHz ADCP are used in this analysis. The integral time scale at different depths is estimated based on the autocorrelation function of the velocity fluctuations and is used to evaluate the mean-square error as a function of the integration time. The implications of these errors in typical ADCP measurements for discharge estimates in natural streams are discussed. Copyright ASCE 2004.