Evaluation of the Dynamic Selection of Exposure Time Method (Dynamic SET) for Mid- and Mean-Section Discharge Measurements

OBERG, KEVIN; DIAZ LOZADA, JOSE M.; GARCIA, CARLOS M.

Fort Collins, Colorado

Conferencia; Hydraulic Measurements & Experimental Methods; 2023

EWRI

Standard methods for measuring streamflow such as those defined by Rantz and others (1982) require that the water velocities be measured for a fixed duration, referred to as the exposure time. In North America, the exposure time is a constant (e.g. 40 s in North America) for all measurement conditions. Not only is this counterintuitive, the rationale for the exposure time is not known according to Pelletier (1988). With widespread adoption of acoustic instruments for making field measurements of velocity, there is an opportunity for re-visiting this topic.The authors have developed a method for dynamically selecting the exposure time for mid- or mean-section discharge measurements (Diaz-Lozada et.al. 2021). This new method makes use of the Moving Block Bootstrap technique (a computational statistical method for uncertainty estimation through re-sampling a time series) to estimate the confidence intervals for measured velocities at each sampling location in the cross-section. In the Dynamic SET method the measured velocity time series is sampled in real time and dynamically determines when the appropriate exposure time has been reached, i.e. when velocity sampling can cease for that particular location. Typically, the appropriate exposure time is selected when the estimated confidence interval is less than a user-specified level of uncertainty. The Dynamic SET method is being implemented in software for collecting mid- and mean-section discharge measurements using acoustic velocity meters, such as ADVs and ADCPs. Currently, the implementation of the method is being evaluated by the authors, the US Geological Survey, and instrument manufacturers. This evaluation consists of determinations of: (1) the maximum exposure time for velocity measurements at each location, (2) the change in reported discharge when the method is used, (3) the change in the time required to make a discharge measurement, and (4) the appropriate uncertainty level that should be used. The use of the coefficient of variation and Standard Deviation of measured velocity to quantify the uncertainty level also are being explored . Preliminary results show that the exposure time recommended for all conditions is not the same and that reported discharge is not substantially different as compared to use of a 40s exposure time. When a large number (>20) of sections (or verticals) are obtained for the discharge measurement, a shorter exposure time can be used because the larger local velocity uncertainties are averaged out.