Streamflow-velocity adjustment coefficients at several common vertical positions varied significantly from station to station and from measurement to measurement at some stations. The ice-cover roughness was the primary factor affecting the profiles at most stations. Several discharge-measurement procedures were developed using one-half of the data collected, and the accuracy of the procedures were evaluated using the other half of the data. The relative error for each method was computed by comparing estimated discharge to discharge computed using complete vertical-profile information. Two procedures - a one-point method with velocity measured at 0.4D (where D is the effective depth of flow) using a site-specific adjustment coefficient and a two-point method using velocity measured at 0.4D and 0.8D - were found to be the most accurate in terms of bias and root-mean-squared error. -from Author