Automated correction of systematic errors in high frequency depth measurements above V-notch weirs using time series decomposition

Hydrologic studies require high-quality streamflow measurements. V-notch weirs provide accurate discharge measurements and when properly equipped produce continuous records of streamflow. However, sensor drift and accumulations of floating woody debris cause systematic errors in depth measurements behind weirs, particularly during periods of base flow between storms. Manual processing of high-frequency streamflow data is often subjective and difficult to reproduce. We developed a method to automatically correct erroneous depth time series to accelerate data processing and promote objectivity and reproducibility. The method inspects depth measurements, isolates and preserves periods of direct runoff during storms, and corrects erroneous inter-event baseflow. We applied experience obtained during manual data processing to implement an automatic correction method based on time series decomposition to produce a consistent time series of stage data that eliminates obvious errors due to sensor drift and clogging by woody debris. The automatic method offers an objective and reproducible procedure capable of efficiently processing 1 year of stage data in seconds. This method promotes reproducibility and may appeal to water resources experts working with large amounts of data especially at sites where distance, cost, or inclement weather prevents regular instrument maintenance.