Laboratory assessment of alternative stream velocity measurement methods.

Abstract

Understanding streamflow in montane watersheds on regional scales is often incomplete due to a lack of data for small-order streams that link precipitation and snowmelt processes to main stem discharge. This data deficiency is attributed to the prohibitive cost of conventional streamflow measurement methods and the remote location of many small streams. Expedient and low-cost streamflow measurement methods used by resource professionals or citizen scientists can provide scientifically useful solutions to this data deficiency. To this end, four current velocity measurement methods were evaluated in a laboratory flume: the surface float, rising body, velocity head rod, and rising air bubble methods. The methods were tested under a range of stream velocities, cross-sectional depths, and streambed substrates. The resulting measurements provide estimates of precision and bias of each method, as well as method-specific insight and calibration formulas. The mean values of the coefficient of variation, a measure of precision, were 10% for the surface float, 10% for the velocity head rod, 14% for the rising body, and 9% for the air bubble method. The values of scaled mean error, a measure of bias, were -8% for the surface float, -4% for the velocity head rod, -1% for the rising body, and 4% for the air bubble. The velocity head rod and surface float methods were the easiest methods to use, providing greater precision at large (> = 0.6 m/s) and small (<0.6 m/s) velocities, respectively. However, the reliance on a velocity ratio for each of these methods can generate inaccuracy in their results. The rising body method is more challenging to execute and of lower precision than the former two methods but provides low bias measurements. The rising air bubble method has a complex instrument assembly that is considered impractical for potential field user groups.